Next Article in Journal
Multi-Objective Majority–Minority Cellular Automata Algorithm for Global and Engineering Design Optimization
Previous Article in Journal
An Improved Iterated Greedy Algorithm for Solving Collaborative Helicopter Rescue Routing Problem with Time Window and Limited Survival Time
Previous Article in Special Issue
A System Design Perspective for Business Growth in a Crowdsourced Data Labeling Practice
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Review

Algorethics in Healthcare: Balancing Innovation and Integrity in AI Development

1
Department of Allied Health Professions, Azienda Ospedaliero-Universitaria Careggi, 50134 Florence, Italy
2
Centre TISP, Istituto Superiore di Sanità, Via Regina Elena 299, 00161 Rome, Italy
3
Azienda Unità Sanitaria Locale Umbria 1, Via Guerriero Guerra 21, 06127 Perugia, Italy
*
Author to whom correspondence should be addressed.
Algorithms 2024, 17(10), 432; https://doi.org/10.3390/a17100432
Submission received: 3 August 2024 / Revised: 19 September 2024 / Accepted: 24 September 2024 / Published: 27 September 2024
(This article belongs to the Collection Feature Papers in Algorithms)

Abstract

:
The rapid advancement of artificial intelligence (AI) technology has catalyzed unprecedented innovation in the healthcare industry, transforming medical practices and patient care. However, this progress brings significant ethical challenges, highlighting the need for a comprehensive exploration of algorethics—the intersection of algorithm design and ethical considerations. This study aimed to conduct a narrative review of reviews in the field of algorethics with specific key questions. The review utilized a standardized checklist for narrative reviews, including the ANDJ Narrative Checklist, to ensure thoroughness and consistency. Searches were performed on PubMed, Scopus, and Google Scholar. The review revealed a growing emphasis on integrating fairness, transparency, and accountability into AI systems, alongside significant progress in ethical AI development. The importance of collaboration between different domains of scientific production, such as social sciences and standardization (like the IEEE), and the development of guidelines is significantly emphasized, with demonstrated direct impact in the health domain. However, gaps persist, particularly in the lack of standardized evaluation methods and the challenges posed by complex sectors like healthcare. The findings underscore the need and importance for robust data governance to prevent biases and highlight the importance of cross-disciplinary collaboration in creating comprehensive ethical frameworks for AI. The field of algorethics has important applications in the health domain, and there is a significant increase in attention, with a focus on addressing issues and seeking both practical and theoretical solutions. Future research should prioritize establishing standardized evaluation practices for AI, fostering interdisciplinary collaboration, developing sector-specific ethical guidelines, exploring AI’s long-term societal impacts, and enhancing ethical training for developers. Continued attention to emerging ethical standards is also crucial for aligning AI technologies with evolving ethical principles.

1. Introduction

1.1. Applications of Artificial Intelligence Algorithms in Healthcare: Innovation and Perspectives

Artificial intelligence (AI) algorithms are revolutionizing the healthcare sector with advanced solutions for diagnosis, treatment, and patient management. In diagnostic settings, AI is extensively used to analyze medical images, including X-rays, computed tomography (CT), magnetic resonance imaging (MRI), positron emission tomography (PET), and ultrasound, enabling precise detection and monitoring of medical conditions [1,2]. Deep learning algorithms, such as convolutional neural networks (CNNs), identify anomalies in medical images, facilitating early diagnosis of diseases like cancer, cardiac conditions, and neurodegenerative disorders [3]. For example, AI can detect suspicious lesions in PET images indicating abnormal metabolic activity [4] or enhance ultrasound images to assess cardiac function or monitor fetal growth [5]. AI also aids in analyzing histological samples by examining digital images of tissue slides to identify abnormal or cancerous cells, speeding up diagnosis and reducing human error [6,7]. In personalized medicine, AI analyzes genomic and clinical data to tailor treatments to individual patient characteristics, predicting drug responses to customize therapies, thus improving clinical outcomes [8,9]. Moreover, AI optimizes hospital resource management through forecasting and optimization techniques, predicting service demand, planning bed allocation, and managing medical staff, which enhances operational efficiency and reduces patient wait times [10,11]. AI supports virtual assistance and automation via chatbots and virtual assistants powered by natural language processing (NLP), providing health information and chronic disease management support, especially in resource-limited settings [12,13]. Finally, AI accelerates pharmaceutical research by using machine learning and predictive modeling to identify new therapeutic compounds, significantly reducing drug development time and costs [14,15]. The integration of AI into healthcare marks a significant technological shift with the potential to improve care quality and patient outcomes [16]. Nonetheless, addressing ethical and regulatory issues is essential to ensure responsible and sustainable use of these technologies [17,18].

1.2. The Role of Algorethics in Modern Healthcare, Starting Point, and Purpose of the Paper

Algorethics is an emerging field focused on embedding ethical principles into algorithm design and use, particularly for artificial intelligence (AI) and automated systems. It aims to ensure these technologies are developed and employed fairly, transparently, and responsibly. Key areas of algorethics [19] include addressing bias and fairness to prevent discrimination, ensuring transparency and explainability to build trust, and safeguarding privacy and security to protect sensitive data. It also involves establishing responsibility and governance for algorithmic decisions and assessing the social impact of these technologies to align them with human values and social justice. In healthcare, algorethics is crucial for promoting fairness, protecting patient privacy, and ensuring accountability for AI-driven decisions [20]. It enhances clinical outcomes by providing accurate and unbiased diagnoses and recommendations and fosters interdisciplinary collaboration to tackle both technical and ethical challenges. Overall, algorethics is vital for ensuring AI technologies are used responsibly in healthcare, improving care quality while safeguarding patient rights and building trust in medical innovations.
The term “algorethics,” originally coined in Italian as “algoretica,” emerged in 2018 with the publication of the book Oracoli. Tra algoretica e algocrazia by Paolo Benanti, a Franciscan friar and professor of moral theology and bioethics at the Pontifical Gregorian University [21]. The “Accademia della Crusca,” Italy’s premier institution for linguistic authority, has since solidified the term within the Italian lexicon. Despite its relatively recent introduction, “algorethics” has rapidly gained international attention, spurring various global initiatives and becoming a key concept in discussions surrounding the ethics of technology and AI. Although it has yet to be officially recognized in English dictionaries, this lack of formal acknowledgment does not detract from its growing influence. The term is increasingly used by scholars, policymakers, and industry leaders to address ethical issues in artificial intelligence and algorithmic decision-making. Its adoption in numerous strategic initiatives highlights its relevance and underscores the need for a deeper understanding of its implications in the technological landscape [22,23,24,25,26,27,28,29].
To appreciate the significance of “algorethics,” it is helpful to consider an overview of various online sources that explore its definition, application, and importance within the context of AI and ethical governance. For instance, the Accademia della Crusca defines “algorethics” as a neologism that merges “algorithm” with “ethics,” emphasizing its role in integrating ethical considerations into algorithmic design and deployment. This emerging field is focused on addressing the moral implications of AI systems [22].
Global policy discussions have also recognized the importance of algorethics. The Rome Call for AI Ethics highlights the need for ethical guidelines in AI development, positioning algorethics as central to international efforts aimed at ensuring responsible AI practices [23]. Similarly, Notre Dame University provides insights into the potential and challenges of applying algorethics to AI, illustrating the complexities of effectively implementing ethical principles in AI technologies [24].
Further conceptual exploration is provided by sources like Paul Wagle’s web overview, which offers a deeper understanding of algorethics as a framework for promoting ethical behavior in algorithmic decision-making [25]. The importance of algorethics in shaping policy and governance structures is underscored by discussions from the Norwegian Institute of International Affairs, which highlight its application in responsible AI governance [26].
Significant academic contributions to the field of algorethics, particularly those by Paolo Benanti, have played a pivotal role in defining and expanding the concept. Benanti’s work, including his book, has been instrumental in gaining scholarly recognition for algorethics [27]. The Rome Call further reinforces the importance of ethical frameworks in AI, integrating algorethics into broader discussions on responsible AI practices [28]. Meanwhile, discussions from Gaywood explore algorethics in the context of AI governance, emphasizing the need for ethical frameworks to guide AI development and calling for diverse stakeholder involvement in AI oversight to ensure transparency, fairness, and accountability [29].
Overall, “algorethics” is increasingly influential across various domains, from global policy and academic research to practical applications in AI governance. Its growing presence in discussions about responsible AI and ethical frameworks underscores its importance in shaping the future of technology.
The purpose of this narrative review is to examine the role of algorethics in the context of artificial intelligence (AI), focusing on its contributions, opportunities, challenges, and recommendations for ethical AI development.
Specific aims:
  • Evaluate Contributions: Categorize and analyze how algorethics is contributing to the ethical development and deployment of AI technologies, highlighting the emerging themes that illustrate its impact.
  • Explore Opportunities and Challenges: Identify the opportunities that algorethics present for improving AI ethics as well as the key challenges that still need to be addressed.
  • Provide Recommendations: Offer recommendations for enhancing algorethics in AI, aiming to better guide the ethical use and governance of AI systems.

2. Materials and Methods

A narrative review of existing reviews was conducted to explore the field of algorethics. To ensure consistency and thoroughness, the review adhered to the ANDJ Narrative Checklist, which can be accessed online [30].

2.1. Search Strategies

The search was based on targeted searches (a) on Pubmed, Scopus, and Google scholar; (b) a properly designed assessment criteria for the study inclusion; (c) an assessment process; and (d) a bias management strategy.
Studies from journals and/or conferences had to be peer reviewed to be included in the process of preselection described below.
The following keyword were used: Algorithmic Ethics, Ethical AI, Responsible AI, Fairness in Algorithms, Bias in AI, AI Transparency, Algorithmic Accountability, AI Governance, Data Ethics, Ethical Machine Learning, AI Ethics Frameworks, Algorithmic Justice, Ethical Decision-Making in AI, Automated Decision Systems, AI Safety, Algorithmic Transparency, Moral Algorithms, AI Regulation, Ethical Design Principles, Algorithmic Governance, Data Privacy, Ethics of Automation, AI Impact Assessment, Ethical Data Use, Supervised Learning, Unsupervised Learning, Machine Learning, Algorithmic Bias Mitigation. Table 1 reports the key targeted searches.

2.2. Assessment Criteria for the Inclusion

To ensure a rigorous and high-quality narrative review, each selected study was assessed based on the following criteria (see the Supplementary Material for further references):
Clarity of Rationale (N1): This criterion evaluates whether the study clearly articulates the reason for its investigation. The rationale should define the research problem, highlight its significance, and explain why the study is necessary. A well-defined rationale provides context and justifies the research effort. For instance, studies should outline the gap in existing knowledge or practice that the research aims to address, and the relevance of the study to algorithms&ethics.
Design Appropriateness (N2): This criterion assesses whether the study’s design is suitable for answering the research question or hypothesis. The design should align with the objectives and scope of the study. Appropriate design includes selecting the right methodology, sample size, and data collection methods. For example, if the study aims to analyze trends in algorithms&ethics over time, a longitudinal design would be appropriate, whereas a cross-sectional design might be used for a snapshot of current practices.
Methodological Clarity (N3): Methodological clarity refers to the extent to which the study’s methods are described in detail and are replicable. This includes the transparency of procedures for data collection, analysis, and interpretation. The study should provide clear information on how the data were gathered, the tools and techniques used, and how the analysis was conducted. This clarity ensures that the study can be reproduced or critiqued based on the methodology described.
Result Presentation (N4): This criterion evaluates how effectively the study presents its findings. Results should be clearly organized, accurately reported, and appropriately interpreted. The presentation should include relevant tables, figures, and statistical analyses that support the conclusions drawn. The clarity of result presentation allows readers to understand and evaluate the study’s outcomes and their implications for algorithms&ethics.
Justification of Conclusions (N5): This criterion assesses whether the study’s conclusions are supported by its results. The study should provide a logical link between the data presented and the conclusions drawn. It should discuss the implications of the findings, address limitations, and suggest areas for future research. Justification of conclusions ensures that the study’s outcomes are valid and that the conclusions are based on sound evidence.
Disclosure of Conflicts of Interest (N6): Disclosure of conflicts of interest is crucial for assessing the impartiality and credibility of the study. This criterion checks whether the authors have declared any financial, professional, or personal interests that could bias the research. Full disclosure helps in evaluating the objectivity of the study and ensures that the findings are not influenced by external pressures or biases.
The choice of the component elements of this overview was made taking into account the 5 parameters (N1–N5) evaluated with a score from 1 = minimum to 5 = maximum and 1 parameter (N6) with a binary assessment (Yes/No). These parameters have been identified into:
All the selected studies had to have the parameter N6 with “Yes” and the parameters N1–N5 with a score > 3.

2.3. Assessment Process

Each study was reviewed by two initial assessors ([DG], [AL]). These assessors were tasked with evaluating each study based on the focus on algorethics and after with the defined criteria. Each criterion, Clarity of Rationale, Design Appropriateness, Methodological Clarity, Result Presentation, Justification of Conclusions, and Disclosure of Conflicts of Interest, was scored on a predefined scale to provide a quantitative measure of each study’s quality and relevance.
The primary assessors independently reviewed the studies and assigned scores to each parameter, ensuring that each study was evaluated against the same standards. This dual-assessment approach was designed to enhance the reliability of the review by capturing different perspectives and reducing the likelihood of individual bias influencing the evaluation process.
In instances where the two initial assessors disagreed on the scores or the inclusion of a study, a third assessor from the group [AP] and [GL] was brought in to adjudicate. This third-party assessment was critical for resolving conflicts and ensuring that the final decisions were fair and well justified. The involvement of a third assessor helped to balance differing opinions and provided an additional layer of scrutiny to uphold the integrity of the review process.
The multi-assessor approach was implemented to minimize bias and ensure a thorough and balanced evaluation of the literature. By incorporating diverse viewpoints and providing a structured mechanism for resolving disagreements, the review aimed to offer a comprehensive and objective assessment of the studies’ related algorethics in the health domain.

2.4. Managing Bias in the Narrative Review

To ensure the narrative review was objective and rigorous, several strategies were employed to manage and minimize bias throughout the assessment process. Here’s how biases were managed:
Diverse Assessors:
Each study was reviewed by two primary assessors [DG] and [AL] with different backgrounds. The inclusion of assessors from different backgrounds and expertise levels was intended to capture a range of perspectives and reduce the likelihood of individual biases influencing the evaluation process.
Clear Assessment Criteria:
The assessment was based on defined parameters: Clarity of Rationale, Design Appropriateness, Methodological Clarity, Result Presentation, Justification of Conclusions, and Disclosure of Conflicts of Interest. Furthermore, data were presented based on a standardized checklist. By using predefined parameters, the review process reduced the risk of subjective interpretation.
Scoring System:
Each parameter was scored on a scale from 1 to 5, with a binary assessment for the disclosure of conflicts of interest (Yes/No). This quantifiable approach allowed for consistent evaluation across studies and provided a transparent mechanism for comparing study quality.
Independent Review:
The primary assessors independently reviewed the studies and assigned scores without consulting each other initially. This independence helped to ensure that individual judgments were based solely on the studies’ merit and the predefined criteria, minimizing the influence of groupthink or shared biases.
Dispute Resolution:
In cases where the two primary assessors disagreed on scores or the inclusion of a study, a third assessor, either [AP] or [GL] (with different backgrounds), was involved to resolve the dispute. This third-party adjudication aimed to provide an impartial perspective and resolve conflicts fairly. The involvement of a third assessor added an extra layer of scrutiny and balance to the review process.
Structured Mechanism for Disagreements:
The process for resolving disagreements was structured and formalized. The third assessor reviewed the initial evaluations and provided a reasoned judgment to reconcile differences. This structured approach ensured that conflicts were addressed systematically and that final decisions were based on a comprehensive evaluation.
Transparency:
The use of a standardized checklist for presenting data and a clear scoring system provided transparency in the assessment process. By documenting the criteria and the scoring rationale, the review process was made transparent, allowing for a clear understanding of how decisions were made and reducing the potential for undisclosed biases.
By incorporating these strategies, this review aimed to offer a thorough and balanced evaluation of the literature. The multi-assessor approach, coupled with structured criteria and formal dispute resolution, was designed to minimize bias and enhance the reliability and objectivity of the review process.

2.5. Selected Studies

The procedure ultimately identified studies at the end of the selection process. Figure 1 outlines all the steps involved. The most recent studies produced in the last five years were taken into consideration. Figure 1 illustrates that the initial search yielded a total of 112 reviews. From these, 75 studies were excluded due to their lack of direct focus on algorethics. Following the evaluation according to the methodology described in Section 2.2 and Section 2.3, 15 review studies were retained [31,32,33,34,35,36,37,38,39,40,41,42,43,44,45] for further consideration, while 22 studies were excluded].

2.6. Further Analysis Insights

Given the innovative and complex nature of the emerging topic, the literature gathered from reviews was analyzed and compared not only with documents of national and international relevance but also with studies presented at conferences and symposia. Additionally, research from other relevant databases, including the IEEE database, was incorporated.

3. Results

Section 3.1 presents an editorial overview of the key tabular highlights from the study.
The results have been systematically organized into three subsections. Section 3.2 examines the prevailing trends observed in the studies within this field.
Section 3.3 addresses the specific objectives of the review, focusing on the following key areas:
  • 3.3.1 Assessing the Impact of Algorithmic Ethics: Analyses how algorithmic ethics (algorethics) contributes to the ethical development and deployment of AI technologies. This section categorizes the contributions and highlights emerging themes that demonstrate the impact of algorethics on AI practices.
  • 3.3.2 Identifying Opportunities and Overcoming Challenges in Algorithmic Ethics: Identifies the opportunities that algorethics presents for advancing AI ethics and outlines the major challenges that need to be addressed. This section explores both the potential benefits and the obstacles facing the integration of ethical principles into AI systems.
  • 3.3.3 Strategic Recommendations for Enhancing Algorithmic Ethics: Offers actionable recommendations for improving algorethics in AI. This section aims to guide the ethical use and governance of AI systems, proposing strategies for enhancing ethical practices and addressing identified challenges.
Additionally, a comprehensive analytical summary, with a particular emphasis on the field of algorethics, is provided in the Supplementary Material. This structure is designed to offer a well-organized perspective on the field, presenting detailed insights and practical recommendations for advancing ethical practices in AI.

3.1. Synoptic Diagram of Results

The diagram in Figure 2 provides a highly concise sketch of the results, organized into tabular connections and diagrams, aligned with the overall aim and specific objectives.
Block 1 (from top to bottom) highlights trends that are further supported by the diagram in Figure 3A–C. Block 2 references Table 2, which emphasizes the key areas of interest and focus. Block 3 focuses on the categorization of emerging trends as shown in Table 3. Block 4 reports the practical and emerging implications as reported in Table 4.
Block 5 and Block 6 recall the emerging opportunities and the areas needing broader investigations, as detailed in Table 5 and Table 6, respectively.
Finally, Block 6 recalls the emerging recommendations detailed in Table 7.

3.2. Trends

An analysis of trends in a biomedical database provides valuable insights into how ethical considerations are integrated into algorithmic research, particularly within the healthcare field. We chose PubMed for this analysis, serving as a representative example of broader trends in this field.
The analysis on PubMed, using the combination of the two terms (see Box S1, position 1), identified 495 studies since 1986, out of a total of 286,849 studies focused on algorithms (see Box S1, position 2), representing only 0.17%. This indicates that research addressing the relationship between ethics and algorithms constitutes a relatively small fraction of the overall body of algorithm-focused studies.
When we narrow down to studies that addressed ethical aspects in relation to AI (see Box S1, position 3 in the Supplementary Material), the total number of relevant papers dropped to 78, with the term “algorethics” being coined in 2018. It is particularly interesting to note that the rise in interest in studies focused on AI ethics coincides with the year in which the term “algorethics” was coined by Benanti. It is interesting to analyze the trends both in general ethics studies and those specifically related to AI.
Regarding studies focused on algorithms and ethics, we observed that (Figure 3A) 426 studies were published in the last 10 years, which accounted for 86% of the historical total. When we restricted the search to the last 5 years (overlapping with the COVID-19 pandemic), we saw that 300 studies were published, representing 60.6% of the total.
The data showed a noticeable increase in the volume of studies addressing the ethical implications of algorithms and AI in recent years, particularly during the COVID-19 pandemic. Figure 3B highlights that 77 out of 78 studies in this field were produced over the past 5 years. This surge suggests that the pandemic may have heightened awareness and concern about ethical issues in technology, likely due to the rapid adoption and critical role of algorithms and AI in managing public health responses and information dissemination during the crisis.
The significant increase in research output on algorithmic ethics and AI ethics during the pandemic period underscores a growing recognition of the need to address ethical considerations in these rapidly evolving fields. This trend reflects a broader societal acknowledgment of the importance of integrating ethical frameworks into technological development and application, particularly as technology increasingly impacts various aspects of life during global crises. The timing of the starting and increased interest in AI and ethics in 2018, coinciding with the coining of the term “algoretica” by Benanti [21], further emphasizes the critical importance of addressing ethical dimensions in the development and implementation of AI technologies. Of all 78 studies produced at the intersection of algorithms, ethics, and AI, 15 were reviews (Figure 3C).

3.3. In-Depth Review of Algorithmic Ethics: Evaluating Impact, Opportunities, Challenges, and Strategic Recommendations

This section explores the role of algorithmic ethics in AI development and deployment. It includes an evaluation of its impact, identifies key opportunities and challenges, and provides strategic recommendations for improving ethical practices in AI. This overview aims to offer valuable insights into integrating ethics into algorithmic design and governance.

3.3.1. Assessing the Impact of Algorithmic Ethics

The transformative impact of artificial intelligence (AI) across diverse fields such as pharmacology [36], medical imaging [42], dermatology [33], and education [43] offers both substantial opportunities and significant ethical challenges. While AI technologies promise advancements in efficiency, accuracy, and personalization, their deployment raises critical ethical concerns, including data privacy [32,37], algorithmic bias [31,34], transparency [33,35], and fairness [38]. To develop and use AI systems responsibly, it is essential to establish robust ethical frameworks, promote interdisciplinary collaboration [39], and engage in ongoing research to address these challenges effectively.
Interconnections and Emerging Themes in AI Ethics
1. 
AI Validation and Generalizability:
  • Overview: Effective AI systems must be rigorously validated to ensure their reliability and applicability across diverse scenarios. Jetzmann et al. [31] emphasized the need for thorough validation processes in musculoskeletal ultrasound to confirm the generalizability of AI algorithms. Similarly, Kim et al. [40] highlighted the importance of ongoing evaluation in digital pathology to address validation and interpretability issues.
  • Ethical Implications: Independent and comprehensive validation processes are crucial for ensuring that AI algorithms are unbiased and reliable. This emphasis on validation helps with mitigating potential biases and ensures that AI systems perform consistently across varied real-world conditions, thereby addressing ethical concerns related to algorithmic fairness and reliability.
2. 
Ethical Implications of Data Use:
  • Overview: The ethical use of data is a critical aspect of AI development, involving concerns about data privacy, potential biases, and the necessity for diverse datasets. Daher et al. [32] examined data privacy issues and the risk of biased outcomes in pancreatic cancer detection, while Veritti et al. [37] discussed the implications of data privacy and healthcare inequality. Wang et al. [38] highlighted the importance of addressing biases in medical AI to promote fairness.
  • Ethical Implications: Addressing data privacy and ensuring the representativeness of datasets are essential for developing fair AI systems. These studies underscore the importance of robust ethical guidelines to protect personal information and ensure equitable outcomes in AI applications.
3. 
Algorithmic Bias and Fairness:
  • Overview: Tackling algorithmic bias and ensuring fairness in AI applications are critical to preventing exacerbation of existing disparities. Grybowsky [33] and Maroufi et al. [34] discussed the risks of biased outcomes resulting from skewed training data and emphasized the need for transparent AI systems. Singh et al. [36] also addressed the importance of fairness in pharmacological research.
  • Ethical Implications: Ensuring fairness involves identifying and mitigating biases in AI algorithms, which requires the creation of standardized practices and robust evaluation methods. These efforts are vital for maintaining equity and trust in AI technologies, ensuring that all demographic groups are treated fairly.
4. 
Transparency and Explainability:
  • Overview: Transparency and explainability are essential for fostering trust in AI systems. Vo et al. [35] and Saw et al. [42] stressed the need for AI systems to be transparent and their decision-making processes to be explainable. Maroufi et al. [34] also highlighted the need for clear ethical guidelines to ensure responsible AI use.
  • Ethical Implications: Transparent and explainable AI systems are fundamental to ethical AI development. By making AI processes understandable, these studies advocate for practices that enhance user trust and align with ethical standards, ensuring that AI technologies are used responsibly and effectively.
5. 
Interdisciplinary Collaboration and Ethical Frameworks:
  • Overview: Developing comprehensive ethical frameworks for AI requires interdisciplinary collaboration. Kontiainen et al. [39] proposed using access to justice as a framework for AI ethics, while Akgun et al. [43] and Kazim et al. [41] highlighted the need for integrating various perspectives to address ethical and regulatory challenges.
  • Ethical Implications: Combining insights from multiple disciplines helps in crafting holistic ethical guidelines that address the complexities of AI. This collaborative approach is essential for developing robust frameworks that ensure re-sponsible AI development and implementation.
6. 
Impact on Stakeholders and Society:
  • Overview: AI technologies have far-reaching implications for various stakeholders and societal dynamics. Bonnefon [44] and Jalal et al. [45] explored the influence of AI on stakeholders, including its role in emergency radiology and its impact on patient care and radiologist workload.
  • Ethical Implications: Considering the broader societal impact of AI is crucial for ethical development. These studies emphasize the need for responsible AI practices that account for stakeholder impacts and societal outcomes, ensuring that AI technologies contribute positively to society.
Table 2 identifies key areas of interest and highlights the focus on AI and ethical considerations.
Table 2. Key areas of interest and focus of the analyzed studies.
Table 2. Key areas of interest and focus of the analyzed studies.
Focus AreaMini-SummaryFocusReferences
AI Validation and GeneralizabilityEmphasizes the importance of rigorous validation methods to ensure that AI algorithms are reliable and applicable across different scenarios. AI systems must undergo thorough testing to demonstrate their effectiveness and robustness in diverse real-world conditions.The focus is on the need for independent and comprehensive validation processes to avoid biases and ensure that AI algorithms perform reliably and fairly when deployed in real-world settings. This includes using external datasets for validation.Jetzmann et al. [31], Kim et al. [40]
Ethical Implications of Data UseDiscusses the critical ethical concerns related to the use of data in AI systems, including issues of privacy, potential biases in data, and the need for diverse and representative datasets. Ensuring data security and ethical handling of personal information is paramount.The focus is on addressing concerns about data privacy, algorithmic bias, and the necessity of using diverse and representative datasets to enhance fairness and accuracy in AI applications.Daher et al. [32], Veritti et al. [37], Wang et al. [38]
Algorithmic Bias and FairnessHighlights the risks associated with algorithmic bias, which can lead to unfair and unequal outcomes across different population groups. Ensures that AI systems are developed with fairness in mind to avoid exacerbating existing disparities.The focus is on identifying and mitigating biases in AI algorithms to promote fairness and equity. This involves creating standardized practices for AI development and performance evaluation to ensure that all populations are treated equitably.Grybowsky [33], Maroufi et al. [34], Singh et al. [36]
Transparency and ExplainabilityStresses the need for AI systems to be transparent and their decision-making processes to be explainable. This helps build trust among users and stakeholders by making it clear how and why decisions are made by AI systems.The focus is on developing AI systems that are transparent in their operations and that provide clear explanations of their decision-making processes. This enhances trust and ensures that AI technologies are used ethically.Vo et al. [35], Saw et al. [42], Maroufi et al. [34]
Interdisciplinary Collaboration and Ethical FrameworksAdvocates for a collaborative approach to developing ethical guidelines for AI through interdisciplinary research. This includes integrating perspectives from various fields to create comprehensive and robust ethical frameworks.The focus is on fostering interdisciplinary collaboration to develop holistic ethical guidelines and regulatory frameworks that address the complexities of AI ethics. This approach aims to ensure that AI systems are designed and implemented responsibly.Kontiainen et al. [39], Akgun et al. [43], Kazim et al. [41]
Impact on Stakeholders and SocietyExamines the broader societal implications of AI technologies, including their effects on various stakeholders and potential changes in societal dynamics. This includes considerations of how AI can influence healthcare access, quality, and equity.The focus is on assessing how AI technologies impact different stakeholder groups and societal structures. It emphasizes the importance of ethical considerations in the development and deployment of AI to ensure positive societal outcomes.Bonnefon [44], Jalal et al. [45]
Table 3 categorizes studies based on their focus areas within AI techniques and the associated ethical issues in healthcare. This table organizes research according to the specific AI techniques and applications they explore, alongside the ethical considerations they address.
The studies included cover a range of AI techniques/focus, such as deep learning models for musculoskeletal ultrasound analysis [31], ensemble methods for pancreatic cancer detection [32], and various machine learning approaches for drug discovery [36]. Each study highlights significant ethical considerations related to its focus, including challenges related to data privacy, algorithmic bias, transparency, and fairness.
For instance, Jetzmann et al. [31] emphasized the importance of validating AI algorithms across diverse clinical settings to prevent biases, while Daher et al. [32] discussed the ethical implications of data scarcity and privacy in pancreatic cancer management. Grybowsky [33] and Maroufi et al. [34] addressed concerns about algorithmic bias and the need for fairness in AI systems, stressing the importance of transparency.
Vo et al. [35] and Saw et al. [42] focused on the necessity for transparency and explainability in AI applications to foster trust and accountability. The interdisciplinary frameworks proposed by Kontiainen et al. [39] and Akgun et al. [43] offered a holistic approach to developing ethical guidelines and regulatory measures.
Additionally, the broader societal impacts of AI technologies were explored by Bonnefon [44] and Jalal et al. [45], emphasizing the need for responsible development practices that integrate ethical standards and consider societal implications.
This categorization provides a structured view of how different AI techniques intersect with ethical issues, reflecting ongoing efforts to address and mitigate potential challenges in AI implementation.
Table 3. Categorization of studies AI techniques/focus and ethical issues in healthcare.
Table 3. Categorization of studies AI techniques/focus and ethical issues in healthcare.
StudyAI Techniques/FocusEthical Considerations
Jetzmann et al. [31]Deep Learning (DL): Convolutional neural networks (CNNs) for analyzing musculoskeletal ultrasound (MSK US) images. Machine learning (ML): Some studies utilized conventional ML techniques such as support vector machines (SVMs) and random forests.Validation Gaps: The study revealed that while internal cross-validation techniques (like K-fold) were prevalent, none of the studies included external clinical validation. This raises concerns about the generalizability of AI algorithms across diverse clinical settings and populations, potentially leading to algorithmic biases and inaccuracies that could impact patient outcomes.
Daher et al. [32]Advanced AI Techniques: Deep learning models and ensemble methods for processing imaging and biomarker data related to pancreatic cancer detection and management. Support vector machines (SVMs) and random forests were also noted in some applications.Data Scarcity and Privacy: Issues include the lack of comprehensive and diverse datasets, which can introduce biases in AI models and affect diagnostic accuracy. Data privacy and security are critical due to the sensitive nature of medical information. There is a need for robust ethical frameworks to address these concerns and ensure fairness and confidentiality in AI systems.
Grybowsky [33]Machine Learning (ML): Techniques such as decision trees, neural networks, and logistic regression for various medical applications, including dermatology.Bias and Transparency: Ethical challenges include potential biases arising from skewed training data and lack of transparency in AI decision-making processes. The “black-box” nature of many AI systems complicates understanding and trust. Ethical considerations also involve ensuring informed consent and protecting patient data from misuse. Improving data quality and making AI systems more explainable are key to addressing these issues.
Maroufi et al. [34]Machine Learning (ML): Diverse algorithms, including logistic regression and decision trees, applied to preoperative planning and surgical decision-making for pituitary adenoma surgery.Standardization and Fairness: The study highlighted the diversity of AI/ML algorithms and raised questions about the standardization and fairness of these technologies. Ensuring that AI models are rigorously tested for reliability and fairness, and addressing any potential biases, is essential for equitable patient outcomes and trust in AI-assisted surgical decision-making.
Vo et al. [35]General AI Methods: Including deep learning and ensemble learning techniques applied to various aspects of healthcare.Privacy and Equity: Ethical concerns include data privacy issues, particularly in relation to third parties like insurance companies, and the risk of perpetuating existing healthcare disparities due to biased AI systems. Transparency in AI systems and the development of clear regulations are necessary to address these concerns and ensure responsible AI implementation.
Singh et al. [36]AI in Pharmacological Research: Techniques such as deep learning for drug discovery, target identification, and toxicity prediction. Reinforcement learning was also noted for optimizing drug efficacy.Privacy and Bias: Ethical challenges include ensuring data privacy and security, addressing algorithmic biases that may skew drug efficacy predictions, and maintaining transparency in AI-driven research. It is crucial to implement robust ethical frameworks and maintain human oversight to mitigate these issues and ensure the responsible use of AI in pharmacology.
Veritti et al. [37]Various AI Approaches: Machine learning and deep learning techniques used in ophthalmology for diagnostics and treatment planning.Bias and Transparency: Key ethical issues include biases in AI algorithms leading to unequal healthcare outcomes, the “black-box” problem complicating accountability, and data security concerns. Ensuring that AI models are explainable, improving data quality, and ensuring equitable access are crucial to addressing these challenges and preventing increased healthcare inequality.
Wang et al. [38]Medical AI Techniques: Incorporating deep learning and statistical models for analyzing and improving fairness in medical AI applications.Fairness and Bias: Ethical issues include ensuring fairness by addressing data quality and algorithmic biases. The review underscores the importance of interdisciplinary discussions to bridge gaps in understanding and implementing practical measures for fairness in medical AI. Legal, ethical, and technological measures are necessary to promote equitable outcomes.
Kontiainen et al. [39]Interdisciplinary AI Frameworks: Combining legal, social, and technological perspectives to address systemic challenges in AI ethics and governance.Systemic Fairness and Justice: The study proposes using access to justice as a framework for understanding and addressing algorithmic biases and ensuring fair AI governance. Integrating multiple perspectives helps develop comprehensive solutions to the ethical and regulatory challenges posed by AI, fostering justice and fairness.
Kim et al. [40]AI in Digital Pathology: Convolutional neural networks (CNNs) and other image-based AI tools for diagnostic purposes.Validation and Interpretability: Ethical challenges include ensuring the accuracy and interpretability of AI systems in pathology. Transparency in AI decision-making processes is crucial for trust and effective integration into diagnostic workflows. Ongoing development and evaluation of AI tools are needed to address these challenges and ensure reliable performance.
Kazim et al. [41]AI as a Digital Asset: Focus on the ontological nature of algorithms and their role in representing and capturing value.Ontological and Ethical Implications: The study emphasized the need to understand how digital technologies and AI represent value and align with societal ethical standards. Addressing these foundational shifts is crucial for aligning AI technologies with broader ethical and societal norms.
Saw et al. [42]AI Techniques in Medical Imaging: Deep learning and advanced image processing methods for analyzing medical images.Algorithm Reliability and Equity: Key challenges include ensuring the creation of reliable and fair AI algorithms, establishing best practices for data governance, and developing regulatory frameworks that support innovation while protecting patient privacy. Addressing transparency and equitable access to AI technologies is essential for ethical development.
Akgun et al. [43]AI in Education: Adaptive learning systems and personalized learning engines for K-12 education settings.Privacy and Bias: Ethical concerns include privacy issues, algorithmic bias, and the need for transparency in educational AI systems. Integrating ethical considerations into AI applications in education and providing resources for educators and students to understand these aspects are essential for responsible AI use.
Bonnefon [44]AI Cognitive Analogies: AI systems emulating human cognitive processes, such as fast and slow thinking models.Misleading Analogies and Design: Ethical considerations involve the potential for misunderstandings or misuse of AI due to misleading analogies to human cognition. Clear and responsible design, along with accurate communication about AI capabilities and limitations, is crucial for ethical AI development.
Jalal et al. [45]AI in Emergency Radiology: Automated image analysis and diagnostic algorithms for handling increased imaging volumes.Integration and Oversight: Challenges include ensuring AI systems are accurate, fair, and transparent while maintaining necessary human oversight in emergency care. Developing frameworks to balance the benefits of AI with ethical standards is crucial for improving care quality and patient safety.
In the rapidly advancing field of artificial intelligence (AI), distinguishing between practical and theoretical applications is also crucial for understanding the scope and impact of various AI research efforts. Practical applications involve the direct use of AI techniques to solve real-world problems and address specific challenges in settings such as healthcare, education, and industry. These applications focus on implementation, effectiveness, and the real-world issues encountered when deploying AI systems. Practical studies often deal with tangible outcomes, such as improving diagnostic accuracy or enhancing operational efficiency.
Conversely, theoretical applications explore the foundational principles, conceptual frameworks, and ethical considerations that underpin AI technologies. These studies provide essential insights into the broader implications of AI but may not involve direct experimentation or immediate deployment. Theoretical research often addresses conceptual issues such as ethical guidelines, systemic challenges, and the philosophical implications of AI, offering valuable perspectives for informing practical applications.
It is important to note that while theoretical applications are fundamental for shaping the ethical and conceptual landscape of AI, they often represent a smaller portion of the overall research landscape compared to practical applications. Practical applications tend to be more numerous and diverse due to their direct relevance and immediate utility in addressing specific problems and improving real-world processes.
Table 4 categorizes various studies based on their primary focus and provides justifications for whether they fall under practical or theoretical applications. This classification highlights the immediate relevance of practical studies in implementing and evaluating AI technologies, while also recognizing the crucial role of theoretical research in guiding the development and ethical deployment of AI systems.
Table 4. Distinction between practical and theoretical applications and justifications.
Table 4. Distinction between practical and theoretical applications and justifications.
StudyApplication TypeJustification for Application Type
Jetzmann et al. [31]PracticalExamines the use of deep learning (DL) and machine learning (ML) in medical imaging, focusing on real-world challenges such as validation and generalizability in clinical settings.
Daher et al. [32]PracticalInvestigates the application of advanced AI techniques in cancer detection, addressing practical issues like data privacy and the accuracy of diagnostic algorithms.
Grybowsky [33]PracticalApplies ML techniques to dermatology, highlighting practical concerns such as biases in training data and the need for transparency in AI decision-making processes.
Maroufi et al. [34]PracticalFocuses on using ML algorithms in surgical planning, emphasizing practical needs for standardization and fairness in AI models used in clinical decision-making.
Vo et al. [35]PracticalExplores the implementation of AI methods in various healthcare applications, with a focus on practical issues including data privacy, equity, and transparency of AI systems.
Singh et al. [36]PracticalAnalyzes the use of AI in pharmacological research, addressing practical concerns such as data privacy, algorithmic biases, and maintaining transparency in drug discovery processes.
Veritti et al. [37]PracticalInvestigates AI applications in ophthalmology, focusing on practical challenges related to bias, data security, and the transparency of AI models.
Wang et al. [38]PracticalExamines AI techniques for improving fairness in medical applications, with a practical focus on addressing biases and ensuring equitable outcomes in healthcare settings.
Kontiainen et al. [39]TheoreticalProposes theoretical frameworks for AI ethics, integrating perspectives from legal, social, and technological fields to address systemic challenges in AI governance.
Kim et al. [40]PracticalStudies the application of AI in digital pathology, focusing on ensuring the accuracy and interpretability of AI systems in diagnostic processes.
Kazim et al. [41]TheoreticalExplores the ontological aspects of AI, examining how AI technologies represent and capture value and their alignment with broader societal ethical norms.
Saw et al. [42]PracticalAddresses practical issues in medical imaging with AI, focusing on ensuring algorithm reliability, data governance, and the development of regulatory frameworks.
Akgun et al. [43]PracticalExamines the use of AI in educational settings, with a focus on practical concerns such as privacy, algorithmic bias, and the need for transparency in educational tools.
Bonnefon [44]TheoreticalExplores the theoretical implications of AI systems mimicking human cognitive processes, discussing the ethical considerations and potential for misunderstanding or misuse.
Jalal et al. [45]PracticalInvestigates the application of AI in emergency radiology, focusing on practical challenges such as ensuring accuracy, fairness, and maintaining necessary human oversight.

3.3.2. Identifying Opportunities and Overcoming Challenges in Algorithmic Ethics

Table 5 highlights the significant opportunities for advancing the ethical development and deployment of AI technologies. These opportunities are derived from a comprehensive review of recent studies, emphasizing areas where strategic improvements can enhance the responsible use of AI [31,32,33,34,35,36,37,38,39,40,41,42,43,44,45].
A major opportunity identified is the enhancement of fairness and transparency within AI systems. Wang et al. [38] advocated for interdisciplinary collaboration to address fairness in medical AI. They emphasized that integrating insights from various disciplines can bridge gaps and improve the implementation of fairness measures. Similarly, Vo et al. [35] stressed the importance of clear regulations and transparency to manage data privacy and mitigate potential biases in healthcare AI systems.
The need for improved data governance is another critical opportunity. Daher et al. [32] highlighted the importance of comprehensive and diverse datasets to prevent biases, which is crucial for ensuring accurate and equitable diagnostics. Singh et al. [36] also underscored the significance of robust ethical frameworks to handle data privacy and algorithmic biases in pharmacological research, promoting transparency and human oversight.
In the educational sector, Akgun et al. [43] discussed the potential of AI to enhance personalized learning and automate assessments. They noted the necessity of addressing privacy issues and algorithmic bias to ensure that AI applications in education are ethical and equitable. Similarly, Saw et al. [42] called for best practices in data governance and regulatory frameworks for medical imaging AI, which can foster innovation while safeguarding patient privacy.
Digital pathology offers another promising opportunity. Kim et al. [40] pointed out that AI tools can significantly improve diagnostic accuracy, provided that challenges related to algorithm validation and interpretability are addressed. Ensuring transparency in AI decision-making processes is essential for effective integration into clinical workflows.
Overall, Table 5 reflects a consensus on the need for interdisciplinary approaches, enhanced data governance, and robust ethical frameworks to advance the responsible use of AI technologies across various domains.
Table 6 outlines key areas that require further research and the challenges associated with the ethical implementation of AI technologies. Addressing these challenges is crucial for ensuring the responsible and effective use of AI systems [31,32,33,34,35,36,37,38,39,40,41,42,43,44,45].
One significant challenge is the need for external validation of AI algorithms to confirm their generalizability and reliability. Jetzmann et al. [31] highlighted that while internal validation techniques such as K-fold cross-validation are common, external clinical validation is often lacking. This gap raises concerns about the real-world applicability of AI systems. Maroufi et al. [34] also emphasized the need for standardized and rigorously tested AI/ML algorithms in surgical decision-making to ensure fairness and reliability.
Data security and the risk of exacerbating healthcare inequalities are substantial concerns. Veritti et al. [37] identified these issues and called for improved data quality, enhanced security measures, and equitable access to AI technologies. Kontiainen et al. [39] proposed using the concept of access to justice as a framework to address systemic challenges in AI ethics, advocating for interdisciplinary approaches to tackle algorithmic biases and ensure fair governance.
Bias and transparency in AI systems remain critical challenges. Grybowsky [33] stressed the need for high-quality data and explainable AI to build trust and mitigate potential biases. Kim et al. [40] further highlighted the importance of transparency and interpretability in AI tools used in digital pathology to ensure reliable decision-making.
In the context of emergency radiology, Jalal et al. [45] addressed the challenge of integrating AI systems while maintaining accuracy, fairness, and necessary human oversight. Ensuring that AI technologies meet high ethical standards and support improved patient care is essential.
In summary, Table 6 underscores the need for ongoing research and development to address these challenges, ensuring that AI technologies are deployed ethically and effectively across healthcare and other fields.
Table 5. Opportunities for advancing AI ethics.
Table 5. Opportunities for advancing AI ethics.
StudyOpportunities
Jetzmann et al. [31]Enhanced AI Validation: There is a significant opportunity to advance AI validation techniques by incorporating external datasets. This approach will enhance the generalizability and reliability of AI algorithms across diverse clinical environments. Effective external validation can ensure that AI models are not limited to specific datasets and can perform reliably in varied real-world settings, thus improving trust in their clinical application.
Daher et al. [32]Diverse Dataset Utilization: Building and utilizing comprehensive datasets that encompass a broad spectrum of demographic and clinical variations represents a key opportunity. By addressing the scarcity of diverse data, AI models can be trained to detect and manage conditions like pancreatic cancer more accurately and fairly. This helps in creating more equitable diagnostic and treatment tools that are effective across different populations.
Grybowsky [33]Increased Transparency: Developing methods and tools to enhance transparency in AI systems can significantly improve stakeholder trust. Creating explainable AI models that clarify decision-making processes can demystify AI operations and build confidence among users, including healthcare professionals and patients. This opportunity focuses on fostering understanding and accountability in AI applications.
Maroufi et al. [34]Standardization of AI Practices: Establishing standardized practices and benchmarks for AI algorithms, especially in surgical contexts, can ensure consistency and fairness in AI-assisted decision-making processes. By setting clear standards for evaluating and implementing AI tools, the medical field can achieve more reliable and equitable outcomes in surgical procedures.
Vo et al. [35]Regulatory Framework Development: Designing and implementing robust regulatory frameworks to address data privacy, equity, and transparency in AI systems presents an opportunity for ethical deployment. Effective regulations can protect patient privacy, ensure fair treatment, and promote transparent AI practices, ultimately fostering a more responsible integration of AI technologies in healthcare.
Singh et al. [36]Ethical Frameworks for Drug Discovery: Creating comprehensive ethical frameworks for AI applications in pharmacology can address bias mitigation, data privacy, and transparency issues. This opportunity involves developing guidelines and standards to ensure that AI-driven drug discovery processes are fair, secure, and transparent, leading to more ethical and effective pharmacological research.
Veritti et al. [37]Improved Data Security and Access: Enhancing data security protocols and ensuring equitable access to AI technologies across patient demographics represent critical opportunities. Improving data protection measures can safeguard sensitive patient information, while ensuring broad access to AI tools can help reduce healthcare disparities and promote fair treatment.
Wang et al. [38]Interdisciplinary Collaboration: Fostering collaboration between different disciplines to address fairness in medical AI offers a chance for more holistic solutions. By integrating perspectives from computer science, medical science, social science, and other fields, researchers can develop comprehensive strategies to address algorithmic biases and improve fairness in AI applications.
Kontiainen et al. [39]Integrated Ethical Frameworks: Utilizing interdisciplinary perspectives to develop comprehensive frameworks for algorithmic fairness and governance can address systemic challenges in AI ethics. This opportunity involves creating integrated ethical guidelines that consider legal, social, and technological aspects, promoting justice and ensuring responsible AI development and use.
Kim et al. [40]Continuous AI Development: Focusing on the ongoing development and refinement of AI tools in digital pathology presents an opportunity to enhance their transparency and integration into diagnostic workflows. Continuous improvement and evaluation of AI systems can ensure their accuracy and reliability, making them valuable tools for pathologists and improving patient outcomes.
Kazim et al. [41]Exploration of Ontological Implications: Delving into the ontological aspects of AI technologies and their value representations can help align AI systems with societal and ethical standards. This opportunity involves examining how AI captures and expresses value, ensuring that its applications are consistent with broader ethical norms and societal expectations.
Saw et al. [42]Best Practices for Data Governance: Developing and implementing best practices for data governance in medical imaging can enhance innovation while safeguarding patient privacy and ensuring transparency. This opportunity involves creating frameworks that balance data protection with the advancement of AI technologies, fostering ethical development and deployment in medical imaging.
Akgun et al. [43]Ethical Education Integration: Integrating ethical considerations into educational resources for AI applications in educational settings represents an opportunity to promote awareness and understanding. By developing instructional materials that address privacy, bias, and transparency, educators and students can navigate the ethical dimensions of AI more effectively.
Bonnefon [44]Clear Design Communication: Ensuring that AI systems are designed and communicated clearly can prevent misunderstandings and misuse. This opportunity involves creating accurate representations of AI capabilities and limitations, promoting responsible design practices and enhancing user understanding of AI technologies.
Jalal et al. [45]Balanced AI Integration: Creating frameworks that balance the benefits of AI with necessary human oversight in emergency radiology can enhance care quality and patient safety. This opportunity focuses on developing guidelines that ensure AI systems complement human expertise while maintaining high standards of care and ethical oversight.
Table 6. Areas needing further research and challenges.
Table 6. Areas needing further research and challenges.
StudyAreas Needing Further Research/Challenges
Jetzmann et al. [31]External Validation: There is a critical need for research on methods to incorporate external validation for AI algorithms in healthcare. Current studies often rely on internal datasets, which may not reflect the variability encountered in real-world clinical settings. To ensure AI models’ generalizability and reliability, research should focus on creating methodologies for effective external validation across diverse patient populations and clinical environments.
Daher et al. [32]Dataset Diversity and Privacy: Addressing the challenges of dataset diversity and data privacy is essential for the ethical deployment of AI. There is a need to develop comprehensive datasets that are representative of different demographic and clinical variations. Additionally, research should focus on enhancing privacy measures to protect sensitive patient data from breaches and on creating ethical guidelines to prevent biases in AI models that could affect diagnostic accuracy.
Grybowsky [33]Bias and Transparency: There is a need for research to identify and mitigate biases in AI algorithms used in medical applications. This includes developing methods to improve transparency in AI decision-making processes to ensure that users can understand and trust AI systems. Investigating ways to make AI systems more explainable and addressing the ethical implications of algorithmic biases are crucial for maintaining fairness and trust in healthcare.
Maroufi et al. [34]Algorithm Selection and Performance: More research is required to determine the best practices for selecting and evaluating AI algorithms in complex scenarios like surgical decision-making. This includes developing standardized criteria for algorithm performance, addressing the diversity of algorithms used, and ensuring that they are tested rigorously for reliability and fairness. Research should also focus on overcoming challenges related to data heterogeneity and algorithmic bias in surgical contexts.
Vo et al. [35]Regulatory Challenges: There is a significant need for research to develop and implement comprehensive regulatory frameworks for AI in healthcare. This includes creating guidelines that address ethical concerns such as data privacy, equity, and transparency. Research should focus on how to balance innovation with ethical considerations and establish clear regulations that ensure AI technologies are used responsibly and effectively in healthcare settings.
Singh et al. [36]Bias and Transparency in Pharmacology: Research should address biases and improve transparency in AI applications within pharmacology. This involves developing methods to identify and mitigate biases in drug discovery and efficacy predictions as well as enhancing transparency in AI-driven research processes. Ensuring data privacy and obtaining informed consent for data use are also critical areas needing further exploration to uphold ethical standards in pharmacological research.
Veritti et al. [37]Data Security and Accessibility: Research should focus on improving data security measures and ensuring equitable access to AI technologies in ophthalmology and healthcare. This includes developing advanced strategies for protecting sensitive health data and addressing the risk of increased healthcare inequality due to biased AI algorithms. Ensuring that AI systems are accessible to diverse patient populations and improving the transparency and security of AI processes are crucial for ethical implementation.
Wang et al. [38]Fairness and Interdisciplinary Approaches: There is a need for more research on interdisciplinary approaches to ensure fairness in medical AI. This involves exploring how various disciplines can collaborate to develop comprehensive solutions for addressing algorithmic biases and promoting fairness. Research should focus on integrating insights from computer science, medical science, and the social sciences to create effective measures for achieving equitable outcomes in AI applications.
Kontiainen et al. [39]Interdisciplinary Ethical Frameworks: Developing integrated ethical frameworks that combine legal, social, and technological perspectives is essential for addressing systemic challenges in AI ethics and governance. Research should focus on creating comprehensive guidelines that address algorithmic biases and ensure justice and fairness in AI systems. Adopting interdisciplinary approaches can help develop more holistic and actionable solutions to the ethical and regulatory challenges posed by AI.
Kim et al. [40]AI Integration in Pathology: Research should focus on the integration of AI tools into digital pathology workflows, emphasizing the need for accuracy, reliability, and transparency in AI systems. Investigating methods to ensure the effective integration of AI into diagnostic processes and continuously developing and refining AI tools to enhance their utility and performance in pathology is crucial for improving diagnostic accuracy and patient outcomes.
Kazim et al. [41]Ontological and Ethical Implications: There is a need to explore the ontological aspects of AI technologies and their ethical implications. Research should investigate how digital technologies represent and process value and ensure that these representations align with broader societal and ethical norms. Understanding the foundational shifts introduced by AI can help ensure that its applications adhere to ethical standards and societal values.
Saw et al. [42]Best Practices for Data Governance: Developing best practices for data governance in medical imaging is essential for ensuring ethical AI development. Research should focus on creating frameworks that balance innovation with privacy protection, transparency, and equitable access. Establishing guidelines for secure data management and transparent AI operations is crucial for fostering ethical development and deployment of AI in healthcare.
Akgun et al. [43]Ethical Education Integration: There is a need to integrate ethical considerations into educational resources for AI applications. This includes developing instructional materials that address privacy, bias, and transparency issues in AI technologies. Promoting awareness and understanding among educators and students about these ethical challenges is essential for preparing future professionals to navigate the ethical implications of AI.
Bonnefon [44]Misleading Analogies and Design: Research should focus on addressing the potential for misunderstandings or misuse of AI due to misleading analogies to human cognition. This involves ensuring that AI systems are designed and communicated in ways that accurately reflect their capabilities and limitations. Clear and responsible design, along with accurate communication about AI systems, is crucial for preventing ethical issues related to cognitive analogies.
Jalal et al. [45]AI in Emergency Radiology: Research should explore the integration of AI in emergency radiology, focusing on challenges such as ensuring accuracy, fairness, and transparency. Developing frameworks that balance the benefits of AI with necessary human oversight and ethical standards is crucial for improving care quality and patient safety in emergency situations. Research should also address how to effectively train and validate AI systems in this high-stakes field.

3.3.3. Strategic Recommendations for Enhancing Algorithmic Ethics

The recommendations outlined in Table 7 are derived from an analysis of the latest research on AI ethics and aim to address key concerns identified across the various studies.
R-1-
Implement Rigorous Validation Practices: The need for comprehensive validation was highlighted by Jetzmann et al. [31] and Maroufi et al. [34]. External validation is essential for ensuring that AI algorithms are not only effective in controlled environments but also in real-world scenarios, reducing the risk of biases and inaccuracies.
R-2-
Develop Comprehensive Data Governance Frameworks: Daher et al. [32] and Singh et al. [36] emphasized the importance of diverse and representative datasets to avoid introducing biases into AI systems. Implementing robust data governance frameworks will help safeguard patient privacy and enhance the fairness and effectiveness of AI technologies.
R-3-
Enhance Transparency and Explainability: As noted by Grybowsky [33], Kim et al. [40], and Jalal et al. [45], transparency and explainability are critical for fostering trust and accountability in AI systems. Making AI decision-making processes more understandable will help address concerns about the “black-box” nature of many AI applications.
R-4-
Promote Fairness through Interdisciplinary Collaboration: Wang et al. [38] and Kontiainen et al. [39] stressed the value of interdisciplinary approaches to address fairness in AI. Collaborative efforts across different fields can lead to more equitable solutions and a better understanding of how fairness can be practically implemented.
R-5-
Address Data Security and Privacy Concerns: Data security and privacy are paramount, especially in sensitive areas like healthcare. Vo et al. [35] and Veritti et al. [37] advocated for the development of clear regulations and guidelines to protect patient information and ensure that AI systems do not perpetuate existing disparities.
R-6-
Integrate Ethical Considerations in Educational AI: Akgun et al. [43] highlighted the need for ethical considerations in AI applications within education. Addressing privacy, bias, and transparency issues will be crucial for ensuring that AI technologies contribute positively to educational outcomes.
R-7-
Advance Research on Ontological Implications: Kazim et al. [41] called for a deeper exploration of the ontological implications of AI technologies. Understanding how AI represents and captures value can help align these technologies with broader societal and ethical standards, ensuring their responsible use.
By addressing these recommendations, stakeholders can work towards a more ethical deployment of AI technologies, enhancing their benefits while mitigating potential risks and challenges.
Table 7. Emerging recommendations for ethical AI deployment.
Table 7. Emerging recommendations for ethical AI deployment.
RecommendationDescriptionReferences
1. Implement Rigorous Validation PracticesEnsure that AI algorithms undergo both internal and external validation to enhance their generalizability and real-world applicability. This includes testing AI systems in diverse clinical settings and populations to verify their reliability and fairness.Jetzmann et al. [31]; Maroufi et al. [34]
2. Develop Comprehensive Data Governance FrameworksCreate and enforce robust ethical frameworks for data collection, usage, and privacy to prevent biases and protect sensitive information. This includes ensuring that datasets are diverse and representative of the populations served.Daher et al. [32]; Singh et al. [36]; Saw et al. [42]
3. Enhance Transparency and ExplainabilityFocus on developing AI systems that are transparent and provide explanations for their decisions. This will help build trust among users and stakeholders and facilitate accountability.Grybowsky [33]; Kim et al. [40]; Jalal et al. [45]
4. Promote Fairness through Interdisciplinary CollaborationFoster interdisciplinary collaborations to address fairness in AI. This includes integrating insights from computer science, medical science, and social science to develop equitable AI systems.Wang et al. [38]; Kontiainen et al. [39]
5. Address Data Security and Privacy ConcernsImplement robust measures to safeguard data security and address privacy concerns, particularly in sensitive areas like healthcare and pharmacology. This includes developing clear regulations and guidelines for data use.Vo et al. [35]; Veritti et al. [37]
6. Integrate Ethical Considerations into Educational AIEnsure that AI applications in education are developed with a focus on ethical considerations such as privacy, bias, and transparency. Provide resources to help educators and students navigate these issues.Akgun et al. [43]
7. Advance Research on Ontological ImplicationsExplore the ontological nature of AI and its role in representing and capturing value. This research should align AI technologies with broader societal and ethical standards.Kazim et al. [41]

4. Discussion

This discussion is organized into five comprehensive subsections, each addressing a critical aspect of the study’s findings and implications.
Section 4.1 provides a synoptic overview that presents an editorial summary of the key tabular highlights from the study. This section organizes and synthesizes the main findings into a clear and accessible format, facilitating a comprehensive understanding of the study’s critical data points and conclusions.
Section 4.2 aligns the study with its initial objectives, critically evaluating how well the outcomes match the predefined goals and objectives. This section emphasizes the added value the research brings to the existing body of knowledge, showcasing how the study enhances understanding and addresses the research questions set at the beginning.
Section 4.3 identifies areas that need further research based on the study’s overview. It highlights specific gaps in the current research and outlines emerging recommendations for future studies. It outlines potential directions for future research, proposing new avenues for inquiry that could build on the current findings and extend the scope of the study. This roadmap for future research aims to address identified deficiencies and explore new opportunities for expanding knowledge in the field.
Section 4.4 extends the discussion by incorporating insights from recent cross-sectional studies across various domains beyond healthcare. It includes an examination of studies from non-biomedical databases, pre-prints, and other relevant sources. This subsection also integrates findings from international documents identified as research priorities through targeted web searches. It is divided into three focused sections: cross-sectional studies, IEEE sources, and international documents. Each section explores how these diverse sources impact the healthcare domain, offering a broad perspective on their relevance and implications.
Finally, Section 4.5 outlines the limitations of the study.

4.1. Synoptic Diagram of Discussion

The diagram in Figure 4 provides a highly concise sketch of the discussion, organized into tabular connections and diagrams, aligned with the overall evolution of the dis-course. Based on the limitations and recommendations that emerged in the overview, Figure 4 elucidates the developments of the discussion through a preliminary analysis of the needs. Block 1 in yellow (from the left to the right) references Table 8 with the insights from cross-disciplinary studies affecting health domain ethics. Block 2 in yellow references Table 9, reporting the IEEE’s perspectives on AI ethics: key findings and implications for healthcare. Connected to Block 1 and Block 2 are two blocks each recalling the section of analysis and of the impact on the health domain. The last block, Block 3 in the center, references Table 10 focusing on the national and international frameworks on algorethics. Also, this block recalls the related section of analysis and the impact on the health domain.

4.2. Discussion on the Added Value of the Review of Reviews and Alignment with the Purpose

A narrative review of existing reviews provides a critical and comprehensive perspective on a field of study, synthesizing multiple findings to offer a broad, integrative view. Unlike traditional reviews, it captures overarching trends and shifts across literature, highlighting emerging patterns, consensus, and disagreements. This synthesis offers insight into the evolution of knowledge, revealing areas of agreement or contention, and helps guide future research.
In the rapidly growing field of algorithmic ethics (algorethics), narrative reviews are particularly important. Algorethics addresses the ethical implications of algorithms and AI systems, such as fairness, transparency, and bias. As AI technologies advance, the need to address these ethical concerns becomes urgent. A narrative review consolidates knowledge, offering an integrated understanding of key issues across diverse domains like medical AI, education, and autonomous systems. This holistic approach helps inform policy and practice while also identifying emerging ethical challenges, such as transparency in AI decision-making and fairness in machine learning models.
Furthermore, such reviews pinpoint gaps in the literature, guiding future research in algorithmic ethics. By synthesizing recent studies, a narrative review provides a timely perspective on pressing issues, contributing to the responsible development and deployment of AI technologies. This review specifically aims to explore algorethics and its implications for AI, addressing key objectives such as evaluating contributions, exploring challenges, and providing recommendations.
Alignment with Research Aims:
Evaluating Contributions:
The review assesses how algorethics has shaped the ethical development of AI across fields like healthcare and education. For instance, Jetzmann et al. [31] emphasized the role of validating AI algorithms in clinical contexts, while Daher et al. [32] discussed data privacy and fairness in AI for cancer detection. Contributions from other studies reinforce the importance of ethical standards in ensuring reliable and unbiased AI technologies.
Exploring Opportunities and Challenges:
Opportunities, such as interdisciplinary collaboration to improve fairness in AI systems, were highlighted by Wang et al. [38]. Challenges, like biases in AI algorithms used in surgery (Maroufi et al. [34]), underscore the need for standardized practices and evaluation methods within algorethics.
Providing Recommendations:
Key recommendations from the review include enhancing transparency and explainability in AI systems, as emphasized by Grybowsky [33] and Kim et al. [40], and developing robust data governance frameworks (Singh et al. [36], Vo et al. [35]) to safeguard privacy and ensure fairness.
This structured approach helps advance knowledge in the evolving field of algorithmic ethics, guiding responsible AI development.

4.3. Discussion: Key Areas for Improvement and Suggested Actions

The value of a narrative review of reviews lies in its ability to highlight established themes within a specific field. This review has demonstrated that studies related to the topic of “algorethics” have only started to emerge from 2018 onward, and the number of reviews on this subject is still limited. A search of PubMed using the terms “algorethics OR algor ethics” (instead of combining the two terms algorithm, ethics with AND as indicated in Box S1, Position 1) returned only one non-review study [46], indicating that the field is still in its developmental stages and relatively embryonic.
Furthermore, this overview also suggests areas that require additional exploration and recommendations. It highlights gaps in the current literature from a stabilization perspective, noting that the volume of review production serves as an indicator of the consolidation and stabilization of themes within a field. As the number and scope of reviews increase, they often signal a maturing and solidifying body of knowledge, reflecting a more stabilized understanding of key issues.
However, despite this progress, there remain critical gaps and evolving ethical considerations in artificial intelligence that necessitate further research. The exploration of “algorethics” reveals several critical areas for improvement in the development and implementation of artificial intelligence (AI).
The identified areas needing broader research/challenges (Table 6) include cross-disciplinary collaboration, algorithmic bias, validation processes, transparency and explainability, data privacy and security, ethical training and education, and long-term societal impacts. Each area presents unique challenges that require targeted strategies to ensure that AI technologies are developed and deployed ethically and responsibly.
The emerging recommendations (Table 7) aim to address key ethical concerns in AI by advocating for rigorous validation practices to ensure real-world effectiveness, developing comprehensive data governance frameworks to prevent biases and protect privacy, and enhancing transparency and explainability to build trust. They also emphasize the importance of interdisciplinary collaboration to promote fairness, addressing data security and privacy concerns, integrating ethical considerations into educational AI, and advancing research on the ontological implications of AI technologies.

4.4. Discussion: Emerging Trends and Contributions in Algorethics from Recent Literature

As the field of artificial intelligence (AI) continues to evolve, ethical considerations have become increasingly crucial. To gain a broader understanding of the current trends and discussions surrounding “algorethics”—the intersection of algorithms and ethics—this review has examined other database sources and documents both to complement the overview and capture the cutting-edge research.
While this review of reviews is predominantly centered on the health domain, there is significant value in exploring scientific literature from other areas that, although not directly related, have considerable impact on the field of AI ethics. For instance, examining research from social sciences and consumer behavior can shed light on broader societal implications of AI technologies. This can be particularly useful in understanding how AI interacts with social norms and consumer expectations, offering insights that may inform ethical considerations in health applications.
Moreover, the technological domain also holds critical relevance. Technological research, particularly in areas responsible for developing and implementing standardization processes, can address many of the emerging recommendations for AI ethics. For example, organizations like the IEEE are actively working on creating standards and guidelines that tackle issues such as transparency, fairness, and accountability in AI systems. These standards can help in aligning ethical practices across various applications of AI, including those in healthcare.
Additionally, integrating insights from preprint articles can provide a forward-looking perspective, capturing the latest trends and emerging concerns before they are widely recognized in formal literature. This approach can help identify new challenges and opportunities for ethical considerations early on.
Furthermore, guidelines and documents produced by national and international bodies play a crucial role in shaping ethical practices. These documents often include comprehensive guidelines that address various aspects of AI ethics, such as data privacy, algorithmic bias, and transparency. By reviewing and incorporating these guidelines, the understanding of ethical issues in AI can be broadened and enriched, leading to more robust and inclusive ethical frameworks.
Overall, while focusing on health-related reviews offers valuable insights, expanding the scope to include literature from other fields, technological standards, and emerging guidelines can provide a more comprehensive understanding of AI ethics. This broader perspective is essential for addressing the complex and evolving ethical challenges associated with artificial intelligence.

4.4.1. Insights from Cross-Disciplinary Studies Affecting Health Domain Ethics

Analysis

Mantini et al. [47] discussed how integrating sustainability principles into AI ethics is vital for ensuring that AI development aligns with long-term societal and environmental goals. Their paper emphasized the need for AI technologies to contribute positively to sustainable development. Benanti et al. [48] highlighted the pressing need for a well-defined ethical framework for AI. The urgency stemmed from the rapid advancement of AI technologies and the corresponding need to address potential ethical issues to ensure responsible development and deployment. Another article from Benanti et al. [49] provided a detailed examination of AI ethics, focusing on the philosophical and practical aspects of implementing ethical standards in AI development. It discussed the implications of AI for ethical theory and practice. Montomoli et al. [50] explored the specific ethical challenges associated with integrating AI into critical care environments. The paper addressed the need for ethical guidelines to navigate the complexities of AI in high-stakes medical situations. Brady et al. [51] reviewed the ethical implications of using AI in radiology, including issues related to data privacy, algorithmic bias, and the impact on clinical decision-making. They underscored the importance of addressing these concerns to ensure the ethical application of AI in healthcare. Anyanwu et al. [52] reflected on the human-centered approach to AI promoted by Pope Francis. The paper discussed how these contributions influence the ethical development of digital technologies and ensure that AI aligns with human values and dignity. Aynla et al. [53] examined how to balance technological progress with ethical considerations in AI. The paper provided practical insights into implementing ethical AI practices while advancing technology. Di Tria et al. [54] addressed the challenges of assessing ethical issues in software, including AI systems. The study proposed methodologies for evaluating the ethical quality of software products and ensuring their responsible use. Amato et al. [55] explored how AI intersects with constitutional values, discussing the implications for legal and ethical standards. The paper examined how AI technologies align with fundamental constitutional principles. Casà et al. [56] proposed a survey assessing the digital skills of young physicians in Italy, highlighting their preparedness for integrating AI technologies into healthcare and addressing the challenges posed by the digital transformation in medicine.
Arokiaswamy et al. [57] discussed the broader economic and social implications of AI, including its impact on employment and social justice. The study examined how AI technologies affect these areas and the ethical considerations necessary to address these impacts.
These references collectively provide a comprehensive view of the current discourse on AI ethics, offering insights into the challenges and frameworks necessary for responsible AI development and implementation. Table 8 reports a sketch of the emerging themes from recent scholarly studies.

Impact on the Health Domain

The integration of AI into the health domain brings substantial ethical implications. Each of the reviewed studies [47,48,49,50,51,52,53,54,55,56,57], although sourced from databases beyond those specifically focusing on the health domain, contributed to understanding how ethical frameworks and practices are crucial for responsible AI deployment in healthcare settings. The following key areas illustrate this impact:
Sustainability and Long-Term Goals.
Mantini et al. [47] argued that incorporating sustainability principles into AI ethics is essential for aligning AI development with long-term societal and environmental goals. In healthcare, this means developing AI technologies that not only advance medical science but also contribute to sustainable healthcare practices. This alignment is crucial for ensuring that AI systems support long-term health outcomes and environmental stewardship.
Urgency of Ethical Frameworks
Benanti et al. [48] emphasized the urgent need for a well-defined ethical framework due to the rapid advancement of AI technologies. The pressing nature of this issue in healthcare is underscored by the need to address potential ethical dilemmas swiftly, ensuring that AI systems are developed and deployed responsibly to avoid unintended consequences such as data breaches or biased diagnostics.
Philosophical and Practical Aspects
Another study by Benanti et al. [49] provided an in-depth examination of AI ethics, focusing on both philosophical and practical aspects. This comprehensive approach helps with understanding the broader implications of AI in healthcare, including how ethical theories and practices can be applied to real-world scenarios, such as patient care and clinical decision-making.
Critical Care Environments
Montomoli et al. [50] explored the specific ethical challenges associated with integrating AI into critical care environments. Their research highlighted the need for tailored ethical guidelines to navigate the complexities of high-stakes medical situations, ensuring that AI applications in critical care are used responsibly and effectively.
Radiology and Data Privacy
Brady et al. [51] reviewed the ethical implications of AI in radiology, addressing issues such as data privacy, algorithmic bias, and the impact on clinical decision-making. Their findings underscore the importance of developing ethical standards to mitigate these concerns, ensuring that AI technologies enhance rather than compromise patient care.
Human-Centered Approach
Anyanwu et al. [52] reflected on the human-centered approach to AI promoted by Pope Francis, discussing how these values influence the ethical development of digital technologies. In healthcare, this approach ensures that AI systems are designed with respect for human dignity and values, promoting patient-centered care.
Balancing Technology and Ethics
Aynla et al. [53] examined how to balance technological progress with ethical considerations in AI. Their insights are valuable for implementing ethical AI practices in healthcare while advancing technology, ensuring that innovation does not come at the expense of ethical standards.
Ethical Quality Assessment
Di Tria et al. [54] addressed the challenges of assessing ethical issues in software, including AI systems. Their proposed methodologies for evaluating the ethical quality of software products are crucial for ensuring that AI systems in healthcare meet high ethical standards and are used responsibly.
Alignment with Constitutional Values
Amato et al. [55] explored the intersection of AI with constitutional values, discussing how AI technologies align with fundamental legal and ethical principles. This alignment is particularly important in healthcare, where AI systems must adhere to principles of justice and equality to ensure fair treatment for all patients.
Digital Skills of Healthcare Professionals
Casà et al. [56] proposed a survey assessing the digital skills of young physicians in Italy, highlighting their preparedness for integrating AI technologies into healthcare. This study underscored the importance of equipping healthcare professionals with the necessary skills to effectively use AI, addressing challenges posed by digital transformation.
Economic and Social Implications
Arokiaswamy et al. [57] discussed the broader economic and social implications of AI, including its impact on employment and social justice. Understanding these implications is crucial for healthcare organizations as they implement AI technologies, ensuring that they address both economic and ethical concerns.
These studies collectively provide a comprehensive view of the current discourse on AI ethics, offering valuable insights into the challenges and frameworks necessary for responsible AI development and implementation in the health domain. The emerging themes from recent scholarly research emphasize the need for robust ethical guidelines, a focus on sustainability, and the importance of preparing healthcare professionals to navigate the complexities of AI technologies effectively.
Table 8. Algorethics insights from recent scholarly cross-disciplinary articles.
Table 8. Algorethics insights from recent scholarly cross-disciplinary articles.
ReferenceFocus on Algorethics and AI
Mantini, A. [47]“Technological Sustainability and Artificial Intelligence Algor-ethics”: Emphasizes integrating sustainability principles into AI ethics to ensure that AI technologies contribute positively to societal and environmental goals.
Benanti, P. [48]“The Urgency of an Algorethics”: Highlights the need for a robust ethical framework for AI due to the rapid advancements in technology and the necessity to address ethical issues proactively.
Benanti, P. [49]“Algor-éthique: Intelligence Artificielle et Réflexion Éthique”: Discusses the integration of ethical standards in AI development, addressing philosophical and practical dimensions of AI ethics.
Montomoli, J., Bitondo, M.M., Cascella, M. [50]“Algor-ethics: Charting the Ethical Path for AI in Critical Care”: Explores the ethical challenges specific to critical care settings, emphasizing the need for tailored ethical guidelines.
Brady, A.P., Neri, E. [51]“Artificial Intelligence in Radiology—Ethical Considerations”: Reviews ethical issues in AI applications in radiology, including data privacy, algorithmic bias, and its impact on clinical decision-making.
Anyanwu, U.S. [52]“Towards a Human-Centered Innovation in Digital Technologies and Artificial Intelligence: The Contributions of the Pontificate of Pope Francis”: Examines how human-centered approaches influence AI ethics, ensuring alignment with human values and dignity.
Aynla et al. [53]“Ethical AI in Practice: Balancing Technological Advancements with Human Values”: Discusses the balance between technological progress and ethical considerations in AI practice.
Di Tria, F. [54]“Measurement of Ethical Issues in Software Products”: Proposes methodologies for assessing the ethical quality of software, including AI systems, to ensure responsible use.
Amato, S. [55]“Artificial Intelligence and Constitutional Values”: Explores how AI technologies must align with constitutional values and fundamental legal and ethical principles.
Casà, C. et al. [56]“COVID-19 and Digital Competencies Among Young Physicians”: Evaluates the digital skills of young physicians in Italy, focusing on their readiness to integrate AI technologies into healthcare.
Arokiaswamy, G. [57]“Artificial Intelligence within the Context of Economy, Employment and Social Justice”: Examines AI’s impact on economy, employment, and social justice, discussing ethical considerations to address these impacts.

4.4.2. IEEE Perspectives on AI Ethics: Key Findings and Implications for Healthcare

Analysis

Examining the intersection of algorithms, AI, and their ethical implications requires a comprehensive look beyond healthcare-specific databases. IEEE publications, including conference papers and standards, offer crucial insights for several reasons. First, the IEEE covers a broad range of topics related to algorithms and ethics, extending beyond any single domain. Conference papers often present cutting-edge research and innovative solutions, capturing a wider view of how ethical considerations and algorithmic advancements are evolving across different fields. This broader scope helps us understand the universal application of ethical principles and provides insights relevant to both healthcare and other domains. Second, IEEE conference proceedings frequently feature the latest research and developments before they appear in journals. These papers address current challenges and emerging trends in real time, ensuring that the review incorporates the most recent advancements in algorithmic ethics and practices. Staying up to date with the latest technologies and ethical standards is crucial for understanding their impact and implications. Third, IEEE sources offer diverse perspectives on algorithmic ethics, including standards, guidelines, and practical case studies [58,59,60,61,62,63,64,65]. For example, the IEEE P7003 Standard for Algorithmic Bias Considerations [62] and the IEEE CertifAIEd Ontological Specification for Algorithmic Bias [65] provide frameworks for addressing biases and ensuring ethical practices in algorithm development. Additionally, educational papers on methods such as role-play case studies [64] present innovative approaches to teaching ethics in technology, essential for preparing future professionals. Moreover, IEEE standards and guidelines offer structured approaches for understanding and addressing ethical issues in algorithms. These frameworks are designed to ensure that algorithms are developed and implemented responsibly, taking into account aspects such as fairness, transparency, and accountability. Including these resources provides a solid foundation for integrating ethical considerations into algorithmic design and deployment. Finally, to fully grasp the impact of algorithms and AI, it is important to consider both technical and ethical dimensions. IEEE sources offer insights into how ethical principles are applied across various scenarios and domains, supporting a holistic understanding of best practices and solutions that address both technical functionality and ethical implications.
As artificial intelligence (AI) continues to transform various sectors, from finance and healthcare to transportation, its impact on society grows increasingly profound. This transformation brings about both significant opportunities and complex ethical challenges. To understand and address these challenges effectively, it is therefore crucial to examine the ethical implications of algorithms and AI through various perspectives. Key areas of concern and advancement in the field have been detected from the IEEE domain, focusing on recent developments, ethical considerations, and the role of standards and frameworks in guiding responsible AI practices.
The Rise and Risks of Algorithms
In the digital age, algorithms permeate nearly every facet of our lives, influencing sectors such as finance, healthcare, and transportation. While their integration offers significant benefits, it also introduces notable risks. The centrality of algorithms in business operations heightens concerns about potential financial losses and reputational damage due to algorithmic failures or biases [58]. This growing concern has given rise to a field dedicated to algorithm auditing aimed at ensuring that AI systems and algorithms are both functional and ethical.
Ethical Challenges in AI Design and Deployment
AI systems are designed to process vast amounts of data and make decisions based on pre-defined rules, which raises substantial ethical challenges. Central to the ethical discourse are principles of fairness, privacy, and accountability [59]. Regulations such as the General Data Protection Regulation (GDPR) impose stringent data protection and privacy requirements, advocating for “privacy by design” in AI systems. This approach seeks to safeguard individual rights while ensuring that AI technologies are deployed responsibly.
Understanding and Addressing Ethical Issues
The deployment of AI across various domains brings about ethical dilemmas, including privacy invasion, algorithmic discrimination, and job displacement [60]. To address these concerns effectively, it is crucial to establish robust ethical guidelines and assess how well AI systems adhere to these standards. The rapidly evolving field of AI ethics is focused on identifying and mitigating the ethical risks associated with AI technologies.
The Role of Standards and Frameworks
To address issues such as algorithmic biases and promote fairness, standards have been developed, such as the IEEE P7003 Standard for Algorithmic Bias Considerations [62]. This standard offers a structured approach for developing algorithms that minimize biases and prevent unjust outcomes. Additionally, the IEEE CertifAIEd Ontological Specification for Algorithmic Bias [65] provides a framework for certifying algorithms with a focus on transparency and fairness, serving as a critical resource for ensuring ethical AI practices.
Moral Decision-Making in AI Projects
The ethical implications of AI extend beyond design and deployment to encompass project management and success criteria [63]. Research underscores the importance of integrating moral decision-making into AI project management, highlighting the need for clear guidelines and best practices. Ensuring that ethical considerations are central to project planning and execution is essential for responsible AI development.
Educating Future Technologists
Preparing the next generation of technologists to navigate the ethical complexities of algorithms is crucial. Educational initiatives, such as role-play case studies [64], are instrumental in enhancing students’ understanding of algorithmic ethics. These methods promote engagement with ethical dilemmas from multiple perspectives, fostering a deeper comprehension of the implications of algorithmic decisions and equipping future professionals with the skills needed to address these challenges.
The Path Forward
As AI technology continues to evolve, the need for robust ethical frameworks, standards, and educational programs becomes increasingly critical [58,59,60,61,62,63,64,65]. Integrating ethics into AI design, deployment, and project management is not only a regulatory necessity but also a moral imperative. Ensuring that AI systems operate in a manner that respects human rights, promotes fairness, and upholds societal values is essential for fostering trust and achieving the transformative potential of these technologies.

Impact on the Health Domain

The influence of IEEE standards and research on the health domain is profound, shaping the development, deployment, and ethical considerations of AI technologies within healthcare. The IEEE’s contributions are pivotal for several reasons:
Establishing Ethical Standards: IEEE standards, such as the IEEE P7003 for Algorithmic Bias Considerations [62], play a crucial role in setting ethical guidelines for AI applications in healthcare. These standards help ensure that AI systems are designed to minimize biases and promote fairness, which is essential for making equitable and accurate medical decisions.
Enhancing Algorithmic Transparency: The IEEE’s frameworks, including the IEEE CertifAIEd Ontological Specification [65], emphasize transparency and accountability in AI systems. This focus on clear, understandable algorithms is particularly important in healthcare, where the transparency of AI decision-making processes can significantly impact trust and patient outcomes.
Guiding Regulatory Compliance: The IEEE’s ethical and regulatory frameworks assist healthcare organizations in meeting legal requirements and industry standards. By adhering to these guidelines, healthcare providers can ensure that AI systems used in clinical settings comply with regulations like the General Data Protection Regulation (GDPR) and contribute to patient privacy and safety [59].
Promoting Innovation and Best Practices: IEEE conferences and publications provide access to cutting-edge research and innovations in AI ethics [58]. This exposure to the latest developments helps healthcare professionals and technologists stay current with best practices, ensuring that AI technologies are implemented effectively and responsibly.
Educating Healthcare Professionals: Educational resources developed by the IEEE, such as role-play case studies [64], are instrumental in preparing healthcare professionals to navigate the ethical challenges of AI. These resources enhance the understanding of algorithmic ethics and support the responsible use of AI technologies in medical practice.
Addressing Practical Challenges: The IEEE’s comprehensive approach to ethical issues provides valuable insights into the practical challenges of integrating AI into healthcare [60]. By addressing both technical and ethical dimensions, IEEE resources help healthcare organizations tackle real-world problems related to AI implementation, such as bias, data privacy, and transparency.
Overall, the IEEE’s standards, frameworks, and educational resources have a significant impact on the health domain by promoting ethical practices, enhancing transparency, guiding regulatory compliance, fostering innovation, and supporting education. These contributions are essential for ensuring that AI technologies are used responsibly and effectively in healthcare settings, ultimately benefiting patient care and safety [65].
Table 9 summarizes the key aspects of each topic related to the ethical implications and standards in algorithmic systems, based on the various IEEE sources provided.
Table 9. Algorethics insights from the IEEE domain.
Table 9. Algorethics insights from the IEEE domain.
TopicDetailsReference
The Rise and Risks of AlgorithmsAlgorithms influence numerous sectors but also pose risks like financial loss and reputational damage. The growing concern over algorithmic failures is driving the development of algorithm auditing to ensure functionality, safety, and ethical considerations.[58]
Ethical Challenges in AI DesignAI systems face ethical issues related to fairness, privacy, and accountability. The GDPR’s “privacy by design” emphasizes data protection and privacy in AI, ensuring that technologies are used responsibly while safeguarding individual rights.[59]
Understanding and Addressing Ethical IssuesEthical dilemmas in AI include privacy invasion, discrimination, and job displacement. Ongoing research aims to identify and mitigate these risks, with a focus on evaluating adherence to ethical standards.[60]
Role of Standards and FrameworksStandards like the IEEE P7003 provide frameworks to address biases and promote fairness in algorithm development. The IEEE CertifAIEd Ontological Specification for Algorithmic Bias offers an ontological approach for certifying algorithms, focusing on transparency and fairness.[62,65]
Moral Decision-Making in AI ProjectsIntegrating moral decision-making into AI project management is crucial. The responsibility for ethical decisions often lies with developers and project managers, highlighting the need for clear guidelines and best practices.[63]
Educating Future TechnologistsEducational methods such as role-play case studies are employed to improve students’ understanding of algorithmic ethics. These approaches help students engage with ethical dilemmas from various perspectives, preparing them to tackle ethical challenges in technology.[64]
The Path ForwardThe evolving nature of AI necessitates robust ethical frameworks, standards, and educational programs. Ensuring AI systems respect human rights and societal values is essential for building trust and realizing the full potential of these technologies.[58,59,60,61,62,63,64,65]

4.4.3. National and International Frameworks on Algorethics

Analysis

The exploration of algorethics has expanded significantly beyond biomedical and scientific databases, extending into various national and international policy frameworks. This extension is reflected in documents produced by prominent organizations such as the World Health Organization (WHO), the European Union (EU), and various national health and regulatory bodies [58,59,60,61,62,63,64]. The focus on algorethics in these documents highlights a growing recognition of the need for ethical standards in AI development and deployment across different sectors.
Global and National Efforts in Algorethics
  • World Health Organization (WHO) [66]: The WHO has released guidelines emphasizing the importance of ethical considerations in the use of AI for global health. These guidelines are part of a broader effort to ensure that AI technologies are developed and applied in ways that respect human rights and promote equity in healthcare. The WHO’s role as an international health authority positions it uniquely to influence global standards and practices in AI ethics.
  • European Union (EU) [67]: The EU’s comprehensive AI Act represents a significant regulatory effort to address ethical concerns related to AI. The act aims to establish a legal framework that ensures AI systems are used responsibly and transparently within the EU. By setting standards for AI risk management and accountability, the EU seeks to balance innovation with ethical responsibility across its member states.
  • FDA (Food and Drug Administration) [68,69]: The FDA has issued guidelines focused on the ethical use of AI in medical research. These guidelines stress the importance of transparency, accountability, and the protection of public health. The FDA’s regulatory oversight ensures that AI technologies in the medical field adhere to high ethical standards, promoting safe and effective use.
  • NHS AI Ethics Initiative [70]: The UK’s NHS AI Ethics Initiative supports the ethical integration of AI in healthcare settings. This initiative provides ethical assurance and manages risks associated with AI technologies, ensuring that healthcare applications of AI maintain a high standard of ethical practice.
  • Public Health Agency of Canada [71]: This document outlines an ethical framework for AI applications in public health. It emphasizes the importance of responsible AI practices and the safeguarding of personal data, reflecting Canada’s commitment to ethical standards in technology deployment.
  • Georgetown University’s Center for Security and Emerging Technologies (CSET) [72]: The CSET document reports ethical norms for AI use in China. The norms covers areas such as the use and protection of personal information, human control over and responsibility for AI, and the avoidance of AI-related monopolies [72]. Table 10 reports a sketch of the national and international documents on algorethics.
Table 10. Overview of national and international documents on algorethics.
Table 10. Overview of national and international documents on algorethics.
ReferenceFocus on AlgorethicsOrganization and RoleExpanded Focus
[66]Global AI Ethics GuidelinesWorld Health Organization (WHO): A leading international public health authority dedicated to addressing global health challenges, including ethical standards for AI technologies, to ensure responsible and equitable development and use.Focuses on global health equity, human rights in AI applications, and responsible AI use across health sectors.
[67]Regulatory Framework for AIEuropean Union (EU): Political and economic union working on the AI Act to set standards for AI use, emphasizing ethical practices, risk management, and accountability across member states.Establishes comprehensive legal and regulatory standards for AI, balancing innovation with ethical considerations and risk management.
[68]Responsible AI Use in Medical ResearchFDA (Food and Drug Administration): Key regulatory body in the United States issuing guidelines for ethical and responsible AI use in medical research, focusing on transparency and public health and safety.Emphasizes transparency, accountability, and safety in AI technologies used in medical research.
[69]Ethical Use of AI in Medical ResearchFDA (Food and Drug Administration): Similar focus as the previous reference, with additional stress on ethical standards in medical research applications.Highlights the importance of ethical guidelines for AI in medical research, ensuring responsible use and public health protection.
[70]AI Ethics in HealthcareNHS (National Health Service) AI Ethics Initiative: Initiative within the UK NHS supporting ethical AI integration in healthcare, providing assurance and managing associated risks.Focuses on ethical integration and risk management of AI in healthcare settings, promoting high standards of ethical practice.
[71]Ethical Framework for AI ApplicationsPublic Health Agency of Canada: Outlines ethical guidelines for AI in public health, emphasizing responsible practices and data protection.Provides a framework for ethical AI use in public health, focusing on responsible practices and personal data protection.
[72]Ethical Norms for AI in ChinaGeorgetown University’s Center for Security and Emerging Technologies (CSET): Research center providing ethical norms for AI in China. The norms cover areas such as the use and protection of personal information, human control over and responsibility for AI, and the avoidance of AI-related monopolies.Focuses on personal information protection and the prevention of monopolistic practices, with limited guidance on enforcement mechanisms.
The integration of algorethics into national and international frameworks reflects a growing commitment to ensuring that AI technologies are developed and used in a manner that upholds ethical standards. By examining the roles and contributions of various organizations, it becomes evident that while significant strides have been made, there remains a need for continued dialogue and refinement of ethical guidelines to address emerging challenges in AI.

Impact on the Health Domain

The integration of AI into the health domain brings profound ethical implications, which are increasingly being addressed through various global and national frameworks. These frameworks, although not always specifically focused on healthcare, significantly impact the health sector by shaping ethical practices and standards for AI technologies. The following key areas illustrate this impact:
Ethical Standards and Equity: The World Health Organization (WHO) has issued global AI ethics guidelines that stress the importance of AI technologies in healthcare respecting human rights and promoting health equity [66]. These guidelines aim to ensure that AI applications are developed and utilized in ways that contribute positively to health outcomes while maintaining ethical standards. The WHO’s leadership in setting these international standards helps drive equitable practices in AI-driven healthcare solutions.
Regulatory Frameworks: The European Union’s AI Act represents a comprehensive effort to regulate AI usage across member states, with a focus on ethical practices, transparency, and accountability [67]. This regulatory framework addresses how AI technologies should be implemented in healthcare settings, striving for a balance between innovation and ethical responsibility. By establishing legal standards for AI, the EU aims to ensure that healthcare applications adhere to high ethical and operational benchmarks.
Transparency and Safety: Guidelines from the U.S. Food and Drug Administration (FDA) on AI in medical research emphasize the need for transparency and safeguarding public health [68,69]. These guidelines are critical for ensuring that AI technologies in healthcare are both effective and ethically deployed. They mandate rigorous standards for transparency, accountability, and safety, fostering trust in AI applications and protecting patient welfare.
Risk Management and Ethical Assurance: The NHS AI Ethics Initiative in the UK supports the ethical integration of AI within healthcare settings by providing assurance and managing associated risks [70]. This initiative focuses on maintaining high standards of ethical practice and addressing potential risks related to AI technologies. By implementing these ethical guidelines, the NHS aims to ensure that AI systems are used responsibly and effectively in clinical environments.
Responsible Practices and Data Protection: The Public Health Agency of Canada has developed an ethical framework for AI applications in public health, which emphasizes responsible practices and the protection of personal data [71]. This framework guides the ethical deployment of AI technologies in healthcare by ensuring that individual privacy is safeguarded while leveraging AI for public health benefits.
Ethical Norms and Enforcement: Georgetown University’s Center for Security and Emerging Technologies (CSET) reports ethical norms for AI use in China. The norms cover areas such as the use and protection of personal information, human control over and responsibility for AI, and the avoidance of AI-related monopolies [72].
Overall, while these documents are not exclusively focused on healthcare, they profoundly influence the health domain by setting ethical principles and standards for AI technologies. They ensure that AI is developed and implemented in ways that are responsible, equitable, and aligned with high ethical standards. This comprehensive approach to algorethics is essential for addressing the complex ethical challenges posed by AI in the healthcare sector.

4.5. Limitations

This review of reviews has some limitations primarily related to its focus, according to the purpose (overview of reviews), on established themes within the field of algorethics. The goal was to synthesize existing knowledge rather than to uncover entirely new perspectives. As a result, some emerging or less consolidated issues might not be completely covered.
However, the overview also incorporated/complemented the analysis with recent non-review studies, conference papers (for example from the IEEE domain), and preprints to offer a current viewpoint on emerging and less consolidated themes. The review also presented an overview of national and international documents. In the case of these documents, the overview primarily utilized publicly accessible documents from national and international sources, which means that some potentially relevant documents that are not available online may be missing.

5. Final Reflections: Broadening Ethical Considerations in New AI Applications beyond Algorithm Development

Recent reviews on AI ethics, especially “algorethics” have predominantly addressed the ethical considerations surrounding algorithm development. While crucial, future discussions must expand to the broader, rapidly evolving ethical landscape of AI applications. For example, Mirzakhani et al. [73] compared AI models with traditional approaches for predicting ICU survival rates, emphasizing the need for robust evaluation protocols beyond algorithms, ensuring real-world applicability and reliability. Wang et al. [74] explored breast cancer prognosis by integrating various risk models, showcasing AI’s potential in personalized medicine and the necessity for responsible integration of these technologies. Lenharo [75] critiqued current AI testing practices in medicine, arguing for rigorous validation to ensure reliability and safety.
The ethical concerns of AI, particularly in healthcare, extend beyond algorithms. Sridhar and Lakshmi [76] addressed privacy, bias, and transparency in AI for diabetes management, emphasizing the importance of ethical frameworks to protect patient rights. Similarly, Fritske et al. [77] explored the ethical challenges of AI-driven polygenic risk scores, calling for new frameworks to address privacy and transparency concerns.
Ethical Sensitivity and Specificity in Healthcare.
Goldberg et al. [78] and Ratwani et al. [79] highlighted the need for ethical sensitivity in deploying AI technologies in healthcare. Goldberg et al. focused on the ethical principles of “do no harm” and “maximize benefit,” stressing the need to carefully balance AI’s potential risks and benefits. Ratwani et al. examined algorithmic bias in healthcare AI systems, advocating for measures like improving dataset diversity and transparency to ensure fairness.
Impact and Challenges of Large Language Models (LLMs).
The rapid development of LLMs has introduced ethical issues, including bias, privacy, misinformation, and transparency [80,81,82]. These systems, like other AI technologies, can perpetuate biases, raising concerns about fairness and privacy. Moreover, LLMs’ capacity to generate content can be misused for spreading misinformation or manipulating public opinion, as Lenharo [75] pointed out.
The emergence of AI applications like Replika [82,83], designed to simulate conversations, has raised concerns about emotional dependency and psychological impact. The Italian privacy authority’s ban on Replika underscores the need for careful evaluation of these technologies. Similarly, “deadbots” [84], which simulate conversations with deceased individuals, raise profound ethical and psychological questions about grief and digital immortality.
In religious contexts, chatbots designed to simulate spiritual conversations, like the recent banned Italian one [85], present unique ethical challenges, potentially affecting sacred traditions and spiritual experiences. The rapid integration of AI into sensitive personal areas reveals a complex web of ethical and psychological issues.
As AI technologies advance, addressing these challenges requires comprehensive ethical frameworks that consider mental health, personal boundaries, and societal values. Transparency, education, and global collaboration among ethicists, technologists, and the public are essential for developing adaptive ethical standards that keep pace with technological advancements.

6. Conclusions and Future Research Directions

6.1. Conclusions

This narrative review of reviews on algorithmic ethics (algorethics) has provided a comprehensive synthesis of emerging themes, opportunities, challenges, and recommendations within this field. By analyzing a broad range of reviews, this study highlights significant advancements and persistent gaps in the ethical considerations surrounding AI. The review underscores the considerable progress made in integrating ethical principles into AI development. It reveals a growing consensus on the importance of embedding fairness, transparency, and accountability into AI systems. These principles are crucial for ensuring that AI technologies do not perpetuate biases or produce unjust outcomes.
Despite these advancements, the review identifies several ongoing challenges. A major issue is the lack of standardized methodologies for evaluating and validating AI systems, which hampers efforts to ensure that these technologies are both effective and ethically sound. The need for robust validation processes is particularly evident in critical applications, such as surgical decision-making, where biases can have serious consequences. Additionally, concerns about data security and the potential for AI to exacerbate inequalities highlight the need for comprehensive ethical frameworks. The review also points to the importance of enhancing cross-disciplinary collaboration. Integrative approaches that combine insights from various fields are essential for developing holistic ethical guidelines that address the diverse applications of AI. Sector-specific ethical guidelines are also needed, as different domains—such as healthcare and education—present unique challenges that require tailored solutions. Overall, while significant progress has been made in the field of algorithmic ethics, addressing existing gaps and challenges remains crucial. This review provides a foundation for advancing the ethical development and deployment of AI technologies by emphasizing the need for standardized evaluation practices, cross-disciplinary collaboration, and sector-specific guidelines.

6.2. Future Research Directions

To address the ongoing challenges in AI ethics, future research should prioritize several key areas. Firstly, it is crucial to establish standardized practices for evaluating and validating AI systems. Consistent methodologies will help ensure that AI technologies are not only effective but also ethically sound, aligning with established ethical principles.
Moreover, fostering cross-disciplinary collaboration is essential for developing comprehensive ethical frameworks. This collaboration should involve experts from diverse fields, including technology, ethics, law, and social sciences, to create a holistic approach that addresses the multifaceted nature of AI applications.
In addition, research should focus on developing sector-specific ethical guidelines tailored to the unique challenges of different domains. For example, exploring the ethical implications of AI in various sectors, such as healthcare and education, can lead to more precise guidelines that ensure the responsible use of AI technologies.
Understanding the long-term societal impacts of AI is also vital. Investigating how AI influences aspects such as employment, privacy, and social equity will help develop strategies to mitigate potential negative consequences and ensure that AI advancements contribute positively to social well-being.
Furthermore, enhancing ethical training and education for technology developers is necessary. As AI becomes increasingly integrated into various aspects of life, it is important that developers are well-versed in ethical principles and their application throughout the design and implementation phases. This includes incorporating ethical considerations into technical curricula and professional development programs.
Finally, staying current with emerging ethical standards and incorporating them into research and development practices will be crucial. The field of AI ethics is rapidly evolving, and adapting to new standards and best practices will help maintain alignment with ethical principles as AI technologies advance.
By focusing on these areas, future research can contribute to the development of AI systems that are both ethically robust and socially responsible.

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/a17100432/s1, Box S1. The proposed composite keys. Supplementary Section S1. Integration of the methodology; Supplementary Section S2. Analytic summaries [86,87,88].

Author Contributions

Conceptualization, D.G.; methodology, D.G.; software, A.L.; validation, All; formal analysis, All; investigation, All; resources, All; data curation, D.G.; writing—original draft preparation, D.G.; writing—review and editing, A.L., A.P., G.L. and D.G.; visualization, All; supervision, D.G.; project administration, D.G. All authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding, APC was funded by Daniele Giansanti.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Shen, D.; Wu, G.; Suk, H.I. Deep Learning in Medical Image Analysis. Annu. Rev. Biomed. Eng. 2017, 19, 221–248. [Google Scholar] [CrossRef] [PubMed]
  2. Li, M.; Jiang, Y.; Zhang, Y.; Zhu, H. Medical image analysis using deep learning algorithms. Front. Public Health 2023, 11, 1273253. [Google Scholar] [CrossRef] [PubMed]
  3. Esteva, A.; Kuprel, B.; Novoa, R.A.; Ko, J.; Swetter, S.M.; Blau, H.M.; Thrun, S. Dermatologist-level classification of skin cancer with deep neural networks. Nature 2017, 542, 115–118. [Google Scholar] [CrossRef] [PubMed]
  4. Visvikis, D.; Lambin, P.; Beuschau Mauridsen, K.; Hustinx, R.; Lassmann, M.; Rischpler, C.; Shi, K.; Pruim, J. Application of artificial intelligence in nuclear medicine and molecular imaging: A review of current status and future perspectives for clinical translation. Eur. J. Nucl. Med. Mol. Imaging 2022, 49, 4452–4463. [Google Scholar] [CrossRef] [PubMed]
  5. Zhang, H.; Qie, Y. Applying Deep Learning to Medical Imaging: A Review. Appl. Sci. 2023, 13, 10521. [Google Scholar] [CrossRef]
  6. Morozov, A.; Taratkin, M.; Bazarkin, A.; Rivas, J.G.; Puliatti, S.; Checcucci, E.; Belenchon, I.R.; Kowalewski, K.F.; Shpikina, A.; Singla, N.; et al. Working Group in Uro-technology of the European Association of Urology. A systematic review and meta-analysis of artificial intelligence diagnostic accuracy in prostate cancer histology identification and grading. Prostate Cancer Prostatic Dis. 2023, 26, 681–692. [Google Scholar] [CrossRef]
  7. Allahqoli, L.; Laganà, A.S.; Mazidimoradi, A.; Salehiniya, H.; Günther, V.; Chiantera, V.; Karimi Goghari, S.; Ghiasvand, M.M.; Rahmani, A.; Momenimovahed, Z.; et al. Diagnosis of Cervical Cancer and Pre-Cancerous Lesions by Artificial Intelligence: A Systematic Review. Diagnostics 2022, 12, 2771. [Google Scholar] [CrossRef]
  8. Vasdev, N.; Gupta, T.; Pawar, B.; Bain, A.; Tekade, R.K. Navigating the future of health care with AI-driven digital therapeutics. Drug Discov. Today 2024, 29, 104110. [Google Scholar] [CrossRef]
  9. Rezayi, S.; RNiakan Kalhori, S.; Saeedi, S. Effectiveness of Artificial Intelligence for Personalized Medicine in Neoplasms: A Systematic Review. BioMed Res. Int. 2022, 2022, 7842566. [Google Scholar] [CrossRef]
  10. Yang, C.C. Explainable Artificial Intelligence for Predictive Modeling in Healthcare. J. Healthc. Inform. Res. 2022, 6, 228–239. [Google Scholar] [CrossRef]
  11. Batko, K.; Ślęzak, A. The use of Big Data Analytics in healthcare. J. Big Data 2022, 9, 3. [Google Scholar] [CrossRef] [PubMed]
  12. Kurniawan, M.H.; Handiyani, H.; Nuraini, T.; Hariyati, R.T.S.; Sutrisno, S. A systematic review of artificial intelligence-powered (AI-powered) chatbot intervention for managing chronic illness. Ann. Med. 2024, 56, 2302980. [Google Scholar] [CrossRef] [PubMed]
  13. Hossain, E.; Rana, R.; Higgins, N.; Soar, J.; Barua, P.D.; Pisani, A.R.; Turner, K. Natural Language Processing in Electronic Health Records in relation to healthcare decision-making: A systematic review. Comput. Biol. Med. 2023, 155, 106649. [Google Scholar] [CrossRef] [PubMed]
  14. Vora, L.K.; Gholap, A.D.; Jetha, K.; Thakur, R.R.S.; Solanki, H.K.; Chavda, V.P. Artificial Intelligence in Pharmaceutical Technology and Drug Delivery Design. Pharmaceutics 2023, 15, 1916. [Google Scholar] [CrossRef] [PubMed] [PubMed Central]
  15. Chen, W.; Liu, X.; Zhang, S.; Chen, S. Artificial intelligence for drug discovery: Resources, methods, and applications. Mol. Ther. Nucleic Acids 2023, 31, 691–702. [Google Scholar] [CrossRef] [PubMed]
  16. Kuziemsky, C.E.; Chrimes, D.; Minshall, S.; Mannerow, M.; Lau, F. AI Quality Standards in Health Care: Rapid Umbrella Review. J. Med. Internet Res. 2024, 26, e54705. [Google Scholar] [CrossRef]
  17. Klimova, B.; Pikhart, M.; Kacetl, J. Ethical issues of the use of AI-driven mobile apps for education. Front. Public Health 2023, 10, 1118116. [Google Scholar] [CrossRef]
  18. Goisauf, M.; Cano Abadía, M. Ethics of AI in Radiology: A Review of Ethical and Societal Implications. Front. Big Data 2022, 5, 850383. [Google Scholar] [CrossRef]
  19. How Ethical AI Transforms Society. Available online: https://algorethics.ai/ (accessed on 20 July 2024).
  20. Oxholm, C.; Christensen, A.S.; Nielsen, A.S. The Ethics of Algorithms in Healthcare. Camb. Q. Healthc. Ethics 2022, 31, 119–130. [Google Scholar] [CrossRef]
  21. Benanti, P. Oracoli. Tra Algoretica e Algocrazia; Luca Sossella Editore: Roma, Italy, 2018. [Google Scholar]
  22. Available online: https://accademiadellacrusca.it/it/parole-nuove/algoretica/18479#:~:text=Parola%20macedonia%20formata%20da%20algor,un%20individuo%20o%20di%20un (accessed on 20 July 2024).
  23. Available online: https://www.romecall.org/algorethics-at-the-un/ (accessed on 20 July 2024).
  24. Available online: https://think.nd.edu/algorethics-potentiality-and-challenges-in-the-age-of-ai/ (accessed on 20 July 2024).
  25. Available online: https://paulwagle.com/what-is-algorethics/ (accessed on 20 July 2024).
  26. Available online: https://www.nupi.no/en/events/2023/algorethics-responsible-governance-of-artificial-intelligence (accessed on 20 July 2024).
  27. Available online: https://www.paolobenanti.com/post/algorethics-oxford (accessed on 20 July 2024).
  28. Available online: https://www.romecall.org/ (accessed on 20 July 2024).
  29. Available online: https://medium.com/@harriet.gaywood/algorethics-who-should-govern-ai-ab1962681078 (accessed on 20 July 2024).
  30. Available online: https://legacyfileshare.elsevier.com/promis_misc/ANDJ%20Narrative%20Review%20Checklist.pdf (accessed on 20 July 2024).
  31. Getzmann, J.M.; Zantonelli, G.; Messina, C.; Albano, D.; Serpi, F.; Gitto, S.; Sconfienza, L.M. The use of artificial intelligence in musculoskeletal ultrasound: A systematic review of the literature. La Radiol. Medica 2024, 129, 1405–1411. [Google Scholar] [CrossRef] [PubMed]
  32. Daher, H.; Punchayil, S.A.; Ismail, A.A.E.; Fernandes, R.R.; Jacob, J.; Algazzar, M.H.; Mansour, M. Advancements in pancreatic cancer detection: Integrating biomarkers, imaging technologies, and machine learning for early diagnosis. Cureus 2024, 16, e56583. [Google Scholar] [CrossRef] [PubMed] [PubMed Central]
  33. Grzybowski, A.; Jin, K.; Wu, H. Challenges of artificial intelligence in medicine and dermatology. Clin. Dermatol. 2024, 42, 210–215. [Google Scholar] [CrossRef] [PubMed]
  34. Maroufi, S.F.; Doğruel, Y.; Pour-Rashidi, A.; Kohli, G.S.; Parker, C.T.; Uchida, T.; Asfour, M.Z.; Martin, C.; Nizzola, M.; De Bonis, A.; et al. Current status of artificial intelligence technologies in pituitary adenoma surgery: A scoping review. Pituitary 2024, 27, 91–128. [Google Scholar] [CrossRef] [PubMed]
  35. Vo, V.; Chen, G.; Aquino, Y.S.J.; Carter, S.M.; Do, Q.N.; Woode, M.E. Multi-stakeholder preferences for the use of artificial intelligence in healthcare: A systematic review and thematic analysis. Soc. Sci. Med. 2023, 338, 116357. [Google Scholar] [CrossRef] [PubMed]
  36. Singh, S.; Kumar, R.; Payra, S.; Singh, S.K. Artificial intelligence and machine learning in pharmacological research: Bridging the gap between data and drug discovery. Cureus 2023, 15, e44359. [Google Scholar] [CrossRef] [PubMed] [PubMed Central]
  37. Veritti, D.; Rubinato, L.; Sarao, V.; De Nardin, A.; Foresti, G.L.; Lanzetta, P. Behind the mask: A critical perspective on the ethical, moral, and legal implications of AI in ophthalmology. Graefe Arch. Clin. Exp. Ophthalmol. 2024, 262, 975–982. [Google Scholar] [CrossRef] [PubMed] [PubMed Central]
  38. Wang, Y.; Song, Y.; Ma, Z.; Han, X. Multidisciplinary considerations of fairness in medical AI: A scoping review. Int. J. Med. Inform. 2023, 178, 105175. [Google Scholar] [CrossRef] [PubMed]
  39. Kontiainen, L.; Koulu, R.; Sankari, S. Research agenda for algorithmic fairness studies: Access to justice lessons for interdisciplinary research. Front. Artif. Intell. 2022, 5, 882134. [Google Scholar] [CrossRef] [PubMed] [PubMed Central]
  40. Kim, I.; Kang, K.; Song, Y.; Kim, T.J. Application of artificial intelligence in pathology: Trends and challenges. Diagnostics 2022, 12, 2794. [Google Scholar] [CrossRef] [PubMed] [PubMed Central]
  41. Kazim, E.; Fenoglio, E.; Hilliard, A.; Koshiyama, A.; Mulligan, C.; Trengove, M.; Gilbert, A.; Gwagwa, A.; Almeida, D.; Godsiff, P.; et al. On the sui generis value capture of new digital technologies: The case of AI. Patterns 2022, 3, 100526. [Google Scholar] [CrossRef] [PubMed] [PubMed Central]
  42. Saw, S.N.; Ng, K.H. Current challenges of implementing artificial intelligence in medical imaging. Phys. Med. 2022, 100, 12–17. [Google Scholar] [CrossRef] [PubMed]
  43. Akgun, S.; Greenhow, C. Artificial intelligence in education: Addressing ethical challenges in K-12 settings. AI Ethics 2022, 2, 431–440. [Google Scholar] [CrossRef] [PubMed] [PubMed Central]
  44. Bonnefon, J.F.; Rahwan, I. Machine Thinking, Fast and Slow. Trends Cogn. Sci. 2020, 24, 1019–1027. [Google Scholar] [CrossRef] [PubMed]
  45. Jalal, S.; Parker, W.; Ferguson, D.; Nicolaou, S. Exploring the role of artificial intelligence in an emergency and trauma radiology department. Can. Assoc. Radiol. J. 2021, 72, 167–174. [Google Scholar] [CrossRef] [PubMed]
  46. Pubmed Search with (Algorethics OR Algor Ethics). Available online: https://pubmed.ncbi.nlm.nih.gov/38573370/ (accessed on 20 July 2024).
  47. Mantini, A. Technological Sustainability and Artificial Intelligence Algor-ethics. Sustainability 2022, 14, 3215. [Google Scholar] [CrossRef]
  48. Benanti, P. The urgency of an algorethics. Discov. Artif. Intell. 2023, 3, 11. [Google Scholar] [CrossRef]
  49. Paolo, B. Algor-éthique: Intelligence artificielle et réflexion éthique. Rev. D’éthique Théologie Morale 2020, 307, 93–110. Available online: https://www.cairn.info/revue-d-ethique-et-de-theologie-morale-2020-3-page-93.htm (accessed on 20 July 2024). [CrossRef]
  50. Montomoli, J.; Bitondo, M.M.; Cascella, M.; Rezoagli, E.; Romeo, L.; Bellini, V.; Semeraro, F.; Gamberini, E.; Frontoni, E.; Agnoletti, V.; et al. Algor-ethics: Charting the ethical path for AI in critical care. J. Clin. Monit. Comput. 2024, 38, 931–939. [Google Scholar] [CrossRef]
  51. Brady, A.P.; Neri, E. Artificial Intelligence in Radiology—Ethical Considerations. Diagnostics 2020, 10, 231. [Google Scholar] [CrossRef]
  52. Anyanwu, U.S. Towards a Human-Centered Innovation in Digital Technologies and Artificial Intelligence: The Contributions of the Pontificate of Pope Francis. Theol. Sci. 2024, 22, 595–613. [Google Scholar] [CrossRef]
  53. Ayinla, B.S.; Amoo, O.O.; Atadoga, A.; Abrahams, T.O.; Osasona, F.; Farayola, O.A. Ethical AI in practice: Balancing technological advancements with human values. Int. J. Sci. Res. Arch. 2024, 11, 1311–1326. [Google Scholar] [CrossRef]
  54. Di Tria, F. Measurement of Ethical Issues in Software Products. Comput. Sci. 2020; preprints. [Google Scholar] [CrossRef]
  55. Amato, S. Artificial Intelligence and Constitutional Values. In Encyclopedia of Contemporary Constitutionalism; Cremades, J., Hermida, C., Eds.; Springer: Cham, Switzerland, 2021. [Google Scholar] [CrossRef]
  56. Casà, C.; Marotta, C.; Di Pumpo, M.; Cozzolino, A.; D’Aviero, A.; Frisicale, E.M.; Silenzi, A.; Gabbrielli, F.; Bertinato, L.; Brusaferro, S. COVID-19 and digital competencies among young physicians: Are we (really) ready for the new era? Ann. Dell’istituto Super. Di Sanita 2021, 57, 1–6. [Google Scholar] [CrossRef] [PubMed]
  57. Arokiaswamy, G. Artificial Intelligence within the Context of Economy, Employment and Social Justice. Asian Horiz. 2020, 14, 628–643. Available online: https://dvkjournals.in/index.php/ah/article/view/3208 (accessed on 20 July 2024).
  58. Koshiyama, A.; Kazim, E.; Treleaven, P. Algorithm Auditing: Managing the Legal, Ethical, and Technological Risks of Artificial Intelligence, Machine Learning, and Associated Algorithms Computer. Computer 2022, 55, 40–50. [Google Scholar] [CrossRef]
  59. Milossi, M.; Alexandropoulou-Egyptiadou, E.; Psannis, K.E. AI Ethics: Algorithmic Determinism or Self-Determination? The GPDR Approach. IEEE Access 2021, 9, 58455–58466. [Google Scholar] [CrossRef]
  60. Huang, C.; Zhang, Z.; Mao, B.; Yao, X. An Overview of Artificial Intelligence Ethics. IEEE Trans. Artif. Intell. 2023, 4, 799–819. [Google Scholar] [CrossRef]
  61. Jameel, T.; Ali, R.; Toheed, I. Ethics of Artificial Intelligence: Research Challenges and Potential Solutions. In Proceedings of the 3rd International Conference on Computing, Mathematics and Engineering Technologies (iCoMET), Lahore, Pakistan, 4–5 March 2020. [Google Scholar]
  62. Koene, A.; Dowthwaite, L.; Seth, S. IEEE P7003TM Standard for Algorithmic Bias Considerations. In Proceedings of the International Workshop on Software Fairness, Kraków, Poland, 12 October 2018. [Google Scholar]
  63. Miller, G.J. Artificial Intelligence Project Success Factors: Moral Decision-Making with Algorithms. In Proceedings of the 16th Conference on Computer Science and Intelligence Systems (FedCSIS), Gdańsk, Poland, 8–11 September 2021. [Google Scholar]
  64. Hingle, A.; Rangwala, H.; Johri, A.; Monea, A. Using Role-Plays to Improve Ethical Understanding of Algorithms Among Computing Students. In Proceedings of the IEEE Frontiers in Education Conference (FIE), Uppsala, Sweden, 13–16 October 2021. [Google Scholar]
  65. IEEE. CertifAIEd™—Ontological Specification for Ethical Algorithmic Bias. Available online: https://engagestandards.ieee.org/rs/211-FYL-955/images/IEEE%20CertifAIEd%20Ontological%20Spec-Algorithmic%20Bias-2022%20%5BI1.3%5D.pdf (accessed on 20 July 2024).
  66. Available online: https://www.who.int/news/item/18-01-2024-who-releases-ai-ethics-and-governance-guidance-for-large-multi-modal-models (accessed on 20 July 2024).
  67. Available online: https://www.modulos.ai/eu-ai-act/?utm_term=ai%20act%20european%20union&utm_campaign=EU+AI+Act+(December+2023)&utm_source=adwords&utm_medium=ppc&hsa_acc=9558976660&hsa_cam=20858946124&hsa_grp=159677877987&hsa_ad=705319461314&hsa_src=g&hsa_tgt=kwd-2178244031979&hsa_kw=ai%20act%20european%20union&hsa_mt=p&hsa_net=adwords&hsa_ver=3&gad_source=1&gclid=CjwKCAjw5Ky1BhAgEiwA5jGujik2Y5RZXOVwXSvUjE-1RARfMpPgen5q2S7-8FnFFLLIiF052SYAwxoC2oEQAvD_BwE (accessed on 20 July 2024).
  68. Available online: https://www.dermatologytimes.com/view/fda-organizations-issue-joint-paper-on-responsible-and-ethical-use-of-artificial-intelligence-in-medical-research (accessed on 20 July 2024).
  69. Available online: https://www.pharmacytimes.com/view/fda-issues-paper-on-the-responsible-use-of-artificial-intelligence-in-medical-research (accessed on 20 July 2024).
  70. Available online: https://transform.england.nhs.uk/ai-lab/ai-lab-programmes/ethics/#:~:text=The%20AI%20Ethics%20Initiative%20supports,risk%20and%20providing%20ethical%20assurance (accessed on 20 July 2024).
  71. Available online: https://www.canada.ca/en/public-health/services/reports-publications/canada-communicable-disease-report-ccdr/monthly-issue/2020-46/issue-6-june-4-2020/ethical-framework-artificial-intelligence-applications.html (accessed on 20 July 2024).
  72. Available online: https://cset.georgetown.edu/publication/ethical-norms-for-new-generation-artificial-intelligence-released (accessed on 20 July 2024).
  73. Mirzakhani, F.; Sadoughi, F.; Hatami, M.; Amirabadizadeh, A. Which model is superior in predicting ICU survival: Artificial intelligence versus conventional approaches. BMC Med. Inform. Decis. Mak. 2022, 22, 167. [Google Scholar] [CrossRef]
  74. Wang, C.; Xu, Y.; Lin, Y.; Zhou, Y.; Mao, F.; Zhang, X.; Shen, S.; Zhang, Y.; Sun, Q. Comparison of CTS5 risk model and 21-gene recurrence score assay in large-scale breast cancer population and combination of CTS5 and recurrence score to develop a novel nomogram for prognosis prediction. Breast 2022, 63, 61–70. [Google Scholar] [CrossRef] [PubMed] [PubMed Central]
  75. Lenharo, M. The testing of AI in medicine is a mess. Here’s how it should be done. Nature 2024, 632, 722–724. [Google Scholar] [CrossRef] [PubMed]
  76. Sridhar, G.R.; Lakshmi, G. Ethical Issues of Artificial Intelligence in Diabetes Mellitus. Med. Res. Arch. 2023, 11. [Google Scholar] [CrossRef]
  77. Fritzsche, M.-C.; Akyüz, K.; Abadía, M.C.; McLennan, S.; Marttinen, P.; Mayrhofer, M.T.; Buyx, A.M. Ethical layering in AI-driven polygenic risk scores—New complexities, new challenges. Front. Genet. 2023, 14, 1098439. [Google Scholar] [CrossRef] [PubMed]
  78. Goldberg, C.B.; Adams, L.; Blumenthal, D.; Brennan, P.F.; Brown, N.; Butte, A.J.; Cheatham, M.; DeBronkart, D.; Dixon, J.; Drazen, J.; et al. To Do No Harm—And the Most Good—With AI in Health Care. Nejm Ai 2024, 1, AIp2400036. [Google Scholar] [CrossRef]
  79. Ratwani, R.M.; Sutton, K.; Galarraga, J.E. Addressing AI Algorithmic Bias in Health Care. JAMA 2024. [Google Scholar] [CrossRef]
  80. Li, J.; Dada, A.; Puladi, B.; Kleesiek, J.; Egger, J. ChatGPT in healthcare: A taxonomy and systematic review. Comput. Methods Programs Biomed. 2024, 245, 108013. [Google Scholar] [CrossRef] [PubMed]
  81. Tian, S.; Jin, Q.; Yeganova, L.; Lai, P.-T.; Zhu, Q.; Chen, X.; Yang, Y.; Chen, Q.; Kim, W.; Comeau, D.C.; et al. Opportunities and challenges for ChatGPT and large language models in biomedicine and health. Briefings Bioinform. 2023, 25, bbad493. [Google Scholar] [CrossRef] [PubMed]
  82. Giansanti, D. The Chatbots Are Invading Us: A Map Point on the Evolution, Applications, Opportunities, and Emerging Problems in the Health Domain. Life 2023, 13, 1130. [Google Scholar] [CrossRef]
  83. Available online: https://apps.apple.com/ch/app/replika-virtual-ai-friend/id1158555867?l=it (accessed on 15 April 2023).
  84. Available online: https://www.cnet.com/culture/hereafter-ai-lets-you-talk-with-your-dead-loved-ones-through-a-chatbot/ (accessed on 15 April 2023).
  85. Available online: https://www.prega.org/ (accessed on 15 April 2023).
  86. Lastrucci, A.; Wandael, Y.; Barra, A.; Ricci, R.; Maccioni, G.; Pirrera, A.; Giansanti, D. Exploring Augmented Reality Integration 307 in Diagnostic Imaging: Myth or Reality? Diagnostics 2024, 14, 1333. [Google Scholar] [CrossRef] [PubMed] [PubMed Central]
  87. Lastrucci, A.; Wandael, Y.; Ricci, R.; Maccioni, G.; Giansanti, D. The Integration of Deep Learning in Radiotherapy: Exploring 310 Challenges, Opportunities, and Future Directions through an Umbrella Review. Diagnostics 2024, 14, 939. [Google Scholar] [CrossRef] [PubMed] [PubMed Central]
  88. Giansanti, D. An Umbrella Review of the Fusion of fMRI and AI in Autism. Diagnostics 2023, 13, 3552. [Google Scholar] [CrossRef] [PubMed] [PubMed Central]
Figure 1. Details on the process of study selection.
Figure 1. Details on the process of study selection.
Algorithms 17 00432 g001
Figure 2. Synoptic diagram presenting an editorial overview of the key tabular highlights from the results.
Figure 2. Synoptic diagram presenting an editorial overview of the key tabular highlights from the results.
Algorithms 17 00432 g002
Figure 3. (A) Publication trends of algorithm and ethics studies: historical vs. recent (last 10 years and last 5 years). (B) Growth of AI ethics research: publications from 2018 to present. (C) Distribution of review articles in AI ethics research (2018–present).
Figure 3. (A) Publication trends of algorithm and ethics studies: historical vs. recent (last 10 years and last 5 years). (B) Growth of AI ethics research: publications from 2018 to present. (C) Distribution of review articles in AI ethics research (2018–present).
Algorithms 17 00432 g003
Figure 4. Synoptic diagram presenting an editorial overview of the key tabular highlights in the discussion.
Figure 4. Synoptic diagram presenting an editorial overview of the key tabular highlights in the discussion.
Algorithms 17 00432 g004
Table 1. The focused search strategy.
Table 1. The focused search strategy.
Keyword FocusKeywordsSearch Query
Algorithmic Ethics and FrameworksAlgorithmic Ethics, Ethical AI, Responsible AI, AI Ethics Frameworks, Ethical Design Principles, Algorithmic Governance, Algorithmic Justice, Algorethics, Algor-Ethics(“Algorithmic Ethics” OR “Ethical AI” OR “Responsible AI” OR “AI Ethics Frameworks” OR “Ethical Design Principles” OR “Algorithmic Governance” OR “Algorithmic Justice” OR “Algorethics” OR “Algor-Ethics”)
Fairness and Bias in AIFairness in Algorithms, Bias in AI, Algorithmic Bias Mitigation, Ethical Machine Learning(“Fairness in Algorithms” OR “Bias in AI” OR “Algorithmic Bias Mitigation” OR “Ethical Machine Learning”)
Transparency and AccountabilityAI Transparency, Algorithmic Accountability, Algorithmic Transparency(“AI Transparency” OR “Algorithmic Accountability” OR “Algorithmic Transparency”)
Governance and RegulationAI Governance, AI Regulation, Ethics of Automation, Data Privacy, Data Ethics(“AI Governance” OR “AI Regulation” OR “Ethics of Automation” OR “Data Privacy” OR “Data Ethics”)
Ethical Decision-Making and SafetyEthical Decision-Making in AI, AI Safety, AI Impact Assessment(“Ethical Decision-Making in AI” OR “AI Safety” OR “AI Impact Assessment”)
Learning MethodsSupervised Learning, Unsupervised Learning, Machine Learning(“Supervised Learning” OR “Unsupervised Learning” OR “Machine Learning”)
Moral and Automated SystemsMoral Algorithms, Automated Decision Systems(“Moral Algorithms” OR “Automated Decision Systems”)
Comprehensive Search QueryAll keywords(“Algorithmic Ethics” OR “Ethical AI” OR “Responsible AI” OR “AI Ethics Frameworks” OR “Ethical Design Principles” OR “Algorithmic Governance” OR “Algorithmic Justice” OR “Algorethics” OR “Algor-Ethics”) AND (“Fairness in Algorithms” OR “Bias in AI” OR “Algorithmic Bias Mitigation” OR “Ethical Machine Learning”) AND (“AI Transparency” OR “Algorithmic Accountability” OR “Algorithmic Transparency”) AND (“AI Governance” OR “AI Regulation” OR “Ethics of Automation” OR “Data Privacy” OR “Data Ethics”) AND (“Ethical Decision-Making in AI” OR “AI Safety” OR “AI Impact Assessment”) AND (“Supervised Learning” OR “Unsupervised Learning” OR “Machine Learning”) AND (“Moral Algorithms” OR “Automated Decision Systems”)
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Lastrucci, A.; Pirrera, A.; Lepri, G.; Giansanti, D. Algorethics in Healthcare: Balancing Innovation and Integrity in AI Development. Algorithms 2024, 17, 432. https://doi.org/10.3390/a17100432

AMA Style

Lastrucci A, Pirrera A, Lepri G, Giansanti D. Algorethics in Healthcare: Balancing Innovation and Integrity in AI Development. Algorithms. 2024; 17(10):432. https://doi.org/10.3390/a17100432

Chicago/Turabian Style

Lastrucci, Andrea, Antonia Pirrera, Graziano Lepri, and Daniele Giansanti. 2024. "Algorethics in Healthcare: Balancing Innovation and Integrity in AI Development" Algorithms 17, no. 10: 432. https://doi.org/10.3390/a17100432

APA Style

Lastrucci, A., Pirrera, A., Lepri, G., & Giansanti, D. (2024). Algorethics in Healthcare: Balancing Innovation and Integrity in AI Development. Algorithms, 17(10), 432. https://doi.org/10.3390/a17100432

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop