Exploring the Intersection of Ergonomics, Design Thinking, and AI/ML in Design Innovation

Abstract

This paper conducts a systematic literature review to explore the dynamic interplay among ergonomics, design thinking, and artificial intelligence (AI) or machine learning (ML). In a rapidly evolving world, the convergence of ergonomics and design Innovation plays a critical role in creating products and spaces valued for their comfort, usability, and aesthetics. Using Elicit, academic papers are identified to elucidate the relationships among these disciplines. The team reviews and selects relevant papers, employing a large language model (LLM)-enabled platform to extract key points. The analysis emphasizes the pivotal roles of ergonomics and design thinking in integrating AI and ML into product design. It underscores the enduring significance of considering user experiences and aesthetics within the AI/ML framework. The findings reveal that while AI/ML techniques enhance precision and innovation in design, integrating ergonomic principles ensures user comfort and safety. The study highlights the necessity for interdisciplinary collaboration and methodological diversity to fully harness the potential of AI and ergonomic design. Limitations include the reliance on contemporary web crawlers and the varying quality of available literature, potentially affecting the comprehensiveness of the review. Future research should focus on developing more robust search methodologies, expanding the scope of studies, and conducting longitudinal research to examine long-term impacts. Ethical implications of AI/ML in design should also be explored to ensure responsible and sustainable use of these technologies. Overall, this research contributes to a nuanced understanding of the roles played by ergonomics and design thinking in synergy with AI/ML for product design, highlighting their impact on shaping user experiences and aesthetics. Despite potential limitations, the study underscores these disciplines’ resilience and lasting relevance in the evolving design field.

1. Introduction

Design thinking and ergonomics are two interdisciplinary fields that play crucial roles in enhancing user experiences and optimizing system performance. Design thinking, a problem-solving approach, emphasizes understanding users, challenging assumptions, and redefining problems to create innovative solutions [1]. On the other hand, ergonomics, also known as human factors, focuses on optimizing interactions between humans and systems to enhance well-being and performance [2].

Design thinking is a multifaceted concept subject to various interpretations and applications across different fields of study and cultural contexts. For example, Berde-Salazer [3] adapted design thinking principles to address wicked problems in social and ecological systems, incorporating insights from various disciplines, while Razzouk and Shute [4] synthesized research to understand design thinking’s characteristics, processes, and potential impact on educational systems, particularly in developing students’ problem-solving skills. Both emphasize that design thinking can be applied in diverse fields, from workplace design to product development, and plays an important role in problem-solving. Understanding the nuances and variations in the definition of design thinking, with the aim of improving user comfort, safety, and efficiency, is essential for contextualizing its relevance and significance within the scope of this study.

The field of ergonomics, also known as “the science of work”, originates from the Greek words ergon (work) and nomos (laws). The terms ergonomics and human factors are often used interchangeably or combined as a unit, such as human factors/ergonomics (HFE or EHF), a convention endorsed by the IEA. The IEA, or the International Ergonomics Association, defines ergonomics (or human factors) as a scientific discipline focused on understanding interactions between humans and various elements within a system. This profession applies theoretical frameworks, principles, data, and methodologies to the design process, aiming to optimize both human well-being and overall system performance. This definition was officially adopted by the IEA in the year 2000 [2].

Moreover, the intersection of these concepts gives rise to the exploration of both “Design Thinking in Ergonomics” and “Ergonomics Design Thinking”, which further enriches our understanding of their synergistic potential in optimizing user-centered design processes. Both “Design Thinking in Ergonomics” and “Ergonomics Design Thinking” can make sense depending on the context and focus of the work:

• “Design Thinking in Ergonomics” suggests that one is applying the principles and methods of design thinking to the field of ergonomics, indicating that one is bringing a design thinking perspective to ergonomic considerations [5].
• “Ergonomics Design Thinking” suggests that one is applying ergonomic principles within the context of design thinking, indicating that one is incorporating ergonomic considerations into the design thinking process [6].

The choice between these two phrases depends on which aspect one wants to emphasize or whether to convey a specific sequence of activities.

While aesthetics and functionality have long been key drivers of innovation, the discipline of ergonomics introduces a critical dimension: the user’s experience [7,8]. The intersection of design thinking and ergonomics presents stirring opportunities for advancing user-centered design processes and optimizing system performance, thereby enhancing user comfort and usability.

Furthermore, as technology continues to advance, the integration of artificial intelligence (AI) or machine learning (ML) into design methodologies becomes increasingly pivotal [9]. AI and ML algorithms offer unprecedented opportunities to analyze vast amounts of data, predict user behaviors, and personalize design solutions. By leveraging AI/ML techniques, designers can refine their understanding of user needs, preferences, and interactions, thus fostering more intuitive and adaptive design processes. This evolution underscores the intersection between cutting-edge technology and human-centric design principles, presenting new frontiers for innovation and problem-solving in various domains [10].

With the continuous evolution of natural language processing (NLP) and machine learning (ML) capabilities, the emergence of large language models (LLMs) such as GPT (generative pre-trained transformer) signifies an innovative advancement in AI-powered applications. These sophisticated models become tools accessed by numerous users in their daily interactions. Their primary function is to emulate human-like language generation and proficiently execute various tasks falling under natural language understanding. In addition to their integration into various domains, LLMs are expected to profoundly impact language-related functionalities and user experiences, shaping the landscape of AI applications [11]. This anticipation is supported by Alqahtani and co-authors work [11], which identified questions covering diverse themes, offering potential avenues for future exploration. This indicates thoughtful consideration of the broader implications, challenges, and potential trade-offs associated with integrating NLP into innovation processes. Six themes are identified: (1) human factor, (2) adoption in organizations, (3) Impact of LLMs, (4) data resources, (5) measuring performance, and (6) research community. Among the identified themes, one stands out as directly connected to a key aspect of our work, the human factor, which concerns potential changes in the skill requirements for individuals involved in innovation, the risks and limitations of depending too much on automated processes, optimizing the collaboration between automated tools and human input, and the impact on human creativity or cognitive aspects as a result of increased reliance on NLP technologies.

In the context of this study, the integration of AI or ML serves a dual purpose. Firstly, AI and ML technologies are utilized to streamline and enhance the literature review process, facilitating a comprehensive examination of the connections between ergonomics and design thinking. Additionally, AI and ML algorithms are employed to investigate how these two disciplines utilize AI and ML techniques themselves. This multifaceted approach enables a deeper understanding of the interplay between ergonomics, design thinking, and advanced technologies, such as AI and ML, in the development of user-centered design solutions and to unveil each discipline’s unique contributions to the others and identify the advantages of algorithmic and ergonomic integration in generative design.

Existing studies have extensively explored the application of HFE in various design activities, highlighting its impact on improving user interaction and system performance [12,13]. However, there remains a significant gap in understanding how the convergence of ergonomics, design thinking, and AI/ML can be leveraged to create innovative, user-centric products. This research aims to address this gap by systematically reviewing relevant literature to elucidate the dynamic interplay among these disciplines and their collective influence on product design.

The research question formulated to address relevant scientific works available and that contribute meaningfully to the field was: “What evidence is there of conjoint use of Ergonomics, Design Thinking, and Artificial Intelligence in product design?”. This research question is focused and clearly states the research objectives defined in the present work.

The introduction section has outlined the background and significance of the study and the research question addressed in this paper. Subsequent sections will delve into the methodology employed, presenting the study design and data collection procedures. Following that, the analysis results will be presented and discussed, providing insights and new trends analysis. Finally, the implications of the findings will be discussed, along with suggestions for future research directions.

2. Materials and Methods

This paper will depict how to lead a systematic literature review (SLR) using different platforms and tools like Elicit as an effective way to gather relevant research papers based on your research question. In the following subsections, a brief review of the meaning of each tool used is presented.

2.1. Elicit and Systematic Literature Review

Elicit is an AI research assistant designed to address the challenges of the literature review process for researchers. It was developed by Ought, a non-profit machine learning lab, and can be accessed through the platform Elicit.org. Elicit utilizes semantic similarity technology in natural language processing (NLP) to generate a list of related papers from Semantic Scholar’s extensive database of 115 million papers. This semantic similarity approach scores text relationships, ordering the search results for user convenience. Elicit limits the results to 400 entries. Notably, Elicit distinguishes itself by employing several large language models (LLMs) to extract relevant metadata, including article title, abstract, authors, citations, and PDF links in the case of open-access articles. This unique feature sets Elicit apart from other scholarly literature search engines, enhancing its utility for literature review researchers [14].

Semantic search is an approach to information retrieval that aims to understand the intent and context behind a user’s query to provide more relevant results. Unlike traditional keyword-based searches that focus on matching specific words or phrases, semantic search considers the meaning of the words and the relationships between them.

Elicit uses the semantic search engine Semantic Scholar, which uses natural language processing and machine learning techniques to comprehend the semantics of the query and the content of documents in their database. This allows for a more sophisticated understanding of user queries, enabling the system to deliver contextually relevant and semantically related results, even if they do not precisely match the query terms. In summary, in the context of a search engine like Semantic Scholar, “semantic” implies a focus on understanding the meaning and context of the information to provide more accurate and contextually relevant search results.

Systematic literature review (SLR), as a research method, involves systematically collecting, analyzing, and synthesizing existing knowledge on a particular topic, providing a comprehensive overview of the current state of research [15]. It helps to identify gaps and contradictions and shed light on emerging trends. This critical analysis informs the direction of future research efforts and contributes to advancing scholarly understanding in a particular domain.

2.2. Research Question Definition

Several steps can be followed when using Elicit as an assistant to locate academic papers for a systematic literature review [14]. The initial step involves defining a research question that should be clear, specific, and aligned with the main objectives of the research and that guides the research process effectively. Based on the search results, this process can be iterative, allowing the refinement and clarification of the search and the quality of the extracted information. That is, in each iteration, the research question is refined based on the preliminary findings. Also, this process was performed collaboratively with all the authors of the present paper, refining and validating the research question. Furthermore, different filters can be considered in the Elicit search, like keywords and publication year, among others.

So, based on the authors’ knowledge and experience, the first research question formulated was: “How do the principles of Ergonomics and Design Thinking synergistically contribute to the effective integration of Artificial Intelligence (AI) and Machine Learning (ML) into product design?”. This question was refined since the observation of the list of existing papers within the domain predominantly addresses education rather than focusing specifically on product design aspects.

Following this first step, the research question considered was: “What evidence is there of conjoint use of Ergonomics, Design Thinking, and Artificial Intelligence in product design?”. The filter ‘Abstract Keywords’ was defined with three keywords: “Artificial Intelligence”, “Human Factors” and “Product Design”, requiring that the abstract of each paper contains the mentioned keywords. To ensure that the search captured the latest research trends and developments, a second filter was considered regarding the ‘Publication year’ targeting works published within the last decade, from 2013 to the present (end 2023). The search was conducted on 15 January.

2.3. Dataset Definition

The search process generates a CSV file containing information from the identified documents, totaling 103 entries. This summary facilitated the authors’ review process, which involved analyzing 102 documents after removing one duplicate.

2.4. Dataset Analysis

The raw data obtained, including the 102 documents, underwent a rigorous analysis involving identifying and categorizing these documents within the predefined scope of the study. The selection process followed a structured approach typical of general literature reviews. This process involves selecting and categorizing all documents according to the flow diagram in Figure 1, which illustrates the data selection and refinement process.

Each document was assessed independently by the authors and assigned one of three categories: “Y” for yes, indicating it should be considered for further analysis; “N” for no, indicating exclusion from further consideration; and “MB” for maybe. The documents marked with uncertainty (MB) were flagged for potential inclusion pending further analysis. This process was essential to ensure the authors’ consensus and refine the dataset. It should be noted that the inclusion criteria were consistently applied, and any discrepancies were resolved through discussions among the reviewing team (authors), emphasizing the reliability and robustness of the data selection process. In this stage of analysis, the inclusion criteria were defined to ensure alignment with the targeted scope of the analysis, encompassing studies directly pertinent to the research topic.

Fifteen documents were selected for further critical analysis since they were considered within the scope of the study and answered the research question defined; eight were considered for potential inclusion, and the remaining were excluded (Figure 1).

In Figure 2, the criteria for the exclusion of documents considered for further analysis are illustrated. This visual representation provides insights into the rigorous selection process employed in our systematic literature review. Documents out of scope were the majority (around 52%), followed by unavailable papers (15%) and missing links to the topic of ergonomics and human factors (6%), among other motives. Notice that the research aims to clarify synergies between ergonomics, design, and artificial intelligence that promote innovation. If one or more of these elements were not addressed concurrently in the papers, they would not align with our analysis goals.

3. Results and Discussion

The ongoing analysis underscores the pivotal roles played by ergonomics and design thinking in integrating AI and ML within product design. The findings highlight the enduring significance of considering user experiences, aesthetics, and the evolving landscape of design practices within the AI/ML framework.

3.1. Bibliographic Network Visualization

VOSviewer v1.6.20 is a software tool for visualizing and analyzing scientific bibliographic data [16]. Visual representations of bibliographic networks, such as co-authorship networks, co-citation networks, and keyword networks, are created to help with a better understanding of the relationships and connections within a scientific field. VOSviewer also includes other features like metrics and analysis tools to assess the impact of authors, journals, or keywords.

Figure 3 illustrates the network generated from bibliographic data, where proximity signifies the strength of co-occurrence links, offering insights into the interconnectedness and relatedness of keywords in the research domain. The co-occurrence analysis, considering author-assigned keywords, was conducted to reveal relationships within the literature. Analyzing keyword co-occurrence unveils trends and guides further research by identifying the most relevant topics and their interconnections.

This network was constructed based on 101 keywords, with a minimum number of one occurrence of a keyword and 62 items with some connection. A minimum occurrence threshold of 2 was applied, focusing on more significant trends and reducing the number to 5 primary keywords: product design, artificial intelligence, machine learning, deep learning, and user experience.

The analysis yielded six distinct clusters, each identified by a different color:

• red cluster: keywords related to “automation”, “interface design”, and “human factors design”,
• green cluster: keywords including “deep learning”, “user experience”, “human intelligence”, “interaction design”, and “machine learning”,
• blue cluster: keywords such as “comfort level”, “ergonomic design”, and “subjective and objective design character”,
• yellow cluster: keywords like “AI ethics”, “artificial intelligence”, and “hybrid intelligence”,
• purple cluster: keywords focusing on “human behavior”, “intelligent systems”, and “sensor”,
• light blue cluster: keywords related to “product design”, “artificial intelligence”, and “chatbot”.

These clusters indicate thematic groupings and highlight the areas where research is densely interconnected, suggesting focal points for ongoing and future research.

It is also possible to analyze the keywords by year, as shown in Figure 4, based on the average occurrence score of the keywords per publication year. This allows us to verify the keywords more recently adopted: “AI ethics”, “chatbot”, and “human behavior”, which occurred after 2022. This reflects the evolving landscape and growing emphasis on ethical considerations, human-centered approaches, and technological advancements in these domains.

The bibliometric analysis provides several key insights essential for understanding the evolving research landscape at the intersection of ergonomics, design thinking, and AI/ML integration. The emergence of new keywords, such as “AI ethics” and “chatbot”, underscores the development of novel trends and expanding areas of interest within this multidisciplinary field. The clustering of keywords further illustrates the strong interdisciplinary connections among ergonomics, design thinking, and AI/ML, suggesting a comprehensive approach to research that considers diverse perspectives and methodologies.

Moreover, identifying less densely populated clusters and infrequently mentioned keywords highlights potential research gaps, offering guidance for future studies to delve into under-explored areas. The temporal analysis reveals shifts in research focus over time, emphasizing the necessity of remaining current with recent advancements and trends. This dynamic evolution reflects the ongoing refinement of methodologies and the increasing importance of addressing contemporary challenges in design and ergonomics.

Overall, the bibliometric analysis using VOSviewer offers a comprehensive overview of the research landscape, aiding scholars and practitioners in understanding the field’s current state and identifying future research directions.

3.2. Thematic Analysis of Selected Papers

After selecting the 15 papers (

Table 1. List and main descriptions of the 15 documents selected (light grey rows identify the five most relevant works identified by Elicit).

Authors/Citation	Data type	Human Factors/Design Target	Methodology	Main Findings
Sung et al. [17]	Quantitative data of 33 right-handed Taiwanese voluntary workers	Anthropometric data according to Grip and key pinch strength of the dominant hand	Regression and artificial neural networks (ANN)	Developed regression models for grip and key pinch strength, with ANN models showing better accuracy.
Wang and Yang [18]	Mix data from a questionnaire survey comparing the innovative image created by designs of the front handle cover with the designs available on the market	Motor scooter’s front handle cover	Three-phase method—preparation, conceptual creativity construction, and semantic analysis	Generated highly innovative concept designs, adaptable for actual products, through comparison with market designs and a survey validation.
Spreafico and Sutrisno [19]	Qualitative data result of 84 specific questions to ask the chatbot, resulting from the combination of known failures and social failures, elements from design theories, and syntactic structures.	Sustainable product design	AI-based chatbot	The proposed method proved its ability to support the social sustainable design of different products and in different ways. However, a dutiful recommendation instead concerns the tool (i.e., the chatbot) due to its filters that limit some answers in which the designer tries to voluntarily hypothesize failures to explore their social consequences.
Alexander and Paul [20]	Qualitative data from literature and internet review	Ergonomics in general	Systematic literature review (SLR)	Identified the need for improvement in DHM systems, especially in the context of emerging ergonomic trends and mobile work system design.
Wang, Choudhary, and Yin [21]	Mixed data: overall product enhancement measured by two keyattributes: product quality and product trendiness	Product quality and product trendiness	Comparative analysis between manual and AI-assisted design	Manual design: deterministic trendiness; AI-assisted design: uncertain trendiness due to the integration of AI technologies. Manual design is preferred when fixed costs are more significant. AI-assisted design is preferred when marginal costs are more crucial. Higher innovation uncertainty favors AI-assisted design.
Ahmed [22]	Qualitative data	Human cognition and smart home usability	Literature review	The psychology of human thought and cognition, everyday things, and designers’ ability to address the role a smart home can play as a leisure equipment/device/object.
Kang [23]	Mixed data	To improve automobile booth design decisions and enhance attendees’ satisfaction	neural network to establish a mapping function between key Kansei factors and representative product design elements	New Miryoku engineering integrating grey relationship analysis, morphological deconstruction, orthogonal design, and neural networks, aims to discover the most perceptually attractive product design
Seabra et al. [24]	Qualitative data	Medical device design	Case study	Highlighted the importance of human-centered design in developing a portable and modular IoT-based detection device for clinical applications. Errors during device operation might hamper effective patient diagnosis and treatment or eventually lead to injury or death.
Felmingham et al. [25]	Qualitative data	Human factors and biases that could arise when using an AI skin cancer diagnostic tool in the real world	Literature review	Emphasized the integration of human factors research in AI diagnostic tool design for skin cancer to enhance effectiveness and address cognitive errors, biases, and unintended consequences.
Nandutu, Atemkeng, and Okouma [26]	Qualitative data	Human and animal behavior	Narrative review, systematic literature review, and bibliometric analysis	Discussed the adoption of AI-based methods for mitigating wildlife–vehicle collisions (WVCs), by understanding and detecting animal behavior of each species.
Whyte [27]	Empirical study	Stakeholders’ reactions to advanced AI products in decision making	Experimental study: considers the impact of AI on decision makingduring cyber conflict episodes in two distinct dimensions	Explored AI’s impact on decision-making during cyber conflicts, highlighting how technical expertise can mitigate biases and improve evaluations of AI inputs.
Mahut et al. [28]	A survey in which 176 participants evaluated 50 interactive products	Interaction design through three dimensions: the environment, the product (target), and the perceived reference (source)	A survey in which 176 participants evaluated 50 interactive products	The study establishes a theoretical model, experiential interaction, linking user experience, interaction, and metaphors. It introduces a tool for classifying interactive products, providing design guidelines based on human affective and cognitive processes.
Duan, Zhang, and Chen [29]	Not explicitly mentioned	How humans interact with products	Uses AR technology and Kansei engineering	Introduced a new KE/DMTM system model combining KE and DMTM to enhance the perceptual and rational constraints in the bionic design of the product.
Mohanty [30]	Mix data: formal and informal interviews and quantitative questionnaire	Complaints regarding the product (office chair); and anthropometric sitting and standing dimensions and a set of design attributes	An integrated approach, combining statistical and artificial intelligence techniques, for ergonomic design and a kinematic model for assessing arm posture	Created an ergonomic office chair that aligns with relevant design standards and offers insights into stress reduction on soft tissues during prolonged sitting.
Geng and Gao [31]	Qualitative data	Maintenance of weapon equipment	Data acquisition module based on fault-tree analysis, utilizing a fault symptom table	Designed a portable fault diagnosis expert system for real-time state acquisition and efficient online diagnosis, promoting intelligence and ease of use for repair and maintenance personnel.

), a comprehensive analysis was performed, delving deep into various aspects to uncover nuanced patterns and insights aligned with the study objective. The authors independently reviewed these papers, collaboratively discussed initial findings, and resolved disagreements through consensus. While reading papers, categories and descriptions, or codes, were identified. These codes, commonly denominated open codes, emerge directly from the data and are not predetermined, allowing for a more exploratory and inductive approach.

The final codes yielded were data type, human factors and/or design target, methodology followed, and main findings obtained. Data type code allows the classification of the type of data analyzed in the selected papers, such as qualitative or quantitative. However, in cases where specific information regarding the data type is not explicitly stated or available within the paper, this information is not provided. Human factors and/or design targets allow for categorizing papers based on their optimization in terms of user experiences, ergonomic characteristics, or design objectives. Methodology as a code allows us to identify the research methodology employed in the selected papers, such as experimental studies, surveys, or SLR. The main findings summarize the key findings or outcomes reported, helping synthesize the collective insights from the 15 selected papers regarding the relationships between ergonomics, design thinking, and AI/ML, including future research directions.

The authors, with their diverse scientific backgrounds and years of experience, played a crucial role in this stage. Each brought unique insights and perspectives, emphasizing their knowledge and expertise in human factors in engineering and technology development, particularly in electronic and computer engineering.

Throughout this process, the author(s) used the information extracted by Elicit, which is available in the columnar view, making it easy for the researcher to synthesize the articles and increasing the researcher’s understanding of the paper, and ChatGPT (https://chat.openai.com/, Version: GPT-3.5) to simply resume the main ideas. Grammarly (https://www.grammarly.com/) was also used to improve readability by providing grammar and spelling suggestions. The authors carefully reviewed and edited the final content.

The order of papers in Table 1 follows Elicit’s selection and relevance accordingly. The rows in light grey on the table highlight the five most relevant works out of the eight identified by Elicit, which the authors also considered crucial within their analysis, contributing significantly to addressing the research question. Of the three works that authors excluded, two were considered out of scope since they were not considered EHF, and the third was not in English (only the title and abstract).

3.2.1. Data Type and Methodology

The analysis of data types and methodologies across the selected studies reveals a rich diversity in approaches, reflecting the multifaceted nature of research at the intersection of ergonomics, design thinking, and AI/ML integration.

Qualitative data was predominantly used in studies such as Sung et al. [17], which utilized anthropometric data to develop regression and artificial neural network (ANN) models for predicting grip and key pinch strength. The robust statistical methods employed here highlight the precision and replicability of quantitative data in ergonomics research. This approach is particularly effective for generating predictive models that inform ergonomic design standards.

Similarly, Mohanty [30] utilized a mixed approach, with formal and informal interviews alongside quantitative questionnaires, to create an ergonomic office chair. Integrating statistical analysis with artificial intelligence techniques showcases the potential of quantitative methods to provide objective, measurable insights into ergonomic design improvements. However, relying on quantitative data alone might overlook nuanced user experiences and subjective comfort factors.

The use of mixed methods was evident in Wang and Yang [18], who combined questionnaire surveys with a three-phase method (preparation, conceptual creativity construction, and semantic analysis) to evaluate motor scooter handle designs. This triangulation of data sources strengthens the validity of their findings, offering a comprehensive view that incorporates both quantitative and qualitative insights. Mixed methods provide a holistic understanding, bridging the gap between numerical data and user perceptions.

Kang [23] also adopted a mixed-methods approach, integrating neural networks to map key Kansei factors to product design elements. By blending qualitative and quantitative data, this study effectively captures the complexity of human factors in design, although the heavy reliance on advanced neural networks may limit the accessibility and reproducibility of the methodology for broader applications.

Studies such as Spreafico and Sutrismo [19] and Felmingham et al. [25] utilized qualitative methodologies. Spreafico and Sutrismo [19] employed qualitative data from chatbot interactions to explore socially sustainable design. While this approach provided rich, context-specific insights, the inherent subjectivity and limitations imposed by the chatbot’s filters may have constrained the depth of the analysis. This highlights a common critique of qualitative methods: the challenge of ensuring consistency and overcoming interpretative biases.

Felmingham et al. [25] conducted a literature review to integrate human factors into the design of AI skin cancer diagnostic tools. The qualitative synthesis offered a nuanced understanding of cognitive errors and biases in clinical settings, underscoring the value of qualitative reviews in identifying complex human-technology interactions. However, the subjective nature of literature reviews can introduce biases based on the reviewers’ perspectives and the selected literature.

Several studies employed systematic literature reviews (SLRs), including Alexander and Paul [20] and Nandutu, Atemkeng, and Okouma [26]. Alexander and Paul [20] employed an SLR to assess digital human modeling (DHM) systems, providing a comprehensive overview of current capabilities and identifying areas for improvement. The structured methodology of SLRs ensures a rigorous, replicable process, although the availability and quality of existing studies may limit it. Nandutu, Atemkeng, and Okouma [26] combined narrative and systematic reviews with bibliometric analysis to discuss AI-based methods for mitigating wildlife–vehicle collisions. This multifaceted approach allowed for a broad yet detailed exploration of the topic, demonstrating the flexibility and depth that systematic reviews can offer. Nonetheless, the effectiveness of SLRs is contingent on the initial search criteria and the databases’ comprehensiveness.

In another study, Wang, Choudhary, and Yin [21] conducted comparative analyses between manual and AI-assisted design strategies, revealing distinct advantages and challenges associated with each approach. This empirical methodology provides clear, actionable insights, though it may oversimplify complex design processes by focusing on binary comparisons.

Whyte [27] employed an experimental study to examine stakeholders’ reactions to AI in decision making during cyber conflicts. The empirical nature of this study provides robust evidence on the impact of technical expertise in mitigating biases. However, experimental designs often face challenges related to ecological validity and the generalizability of findings to real-world settings.

Studies like Duan, Zhang, and Chen [29] and Geng and Gao [31] integrated advanced technologies such as augmented reality (AR) and fault-tree analysis. Duan, Zhang, and Chen [29]’s application of AR and Kansei engineering in product design represents an innovative methodological approach, though the complexity and resource-intensive nature of such technologies may limit their widespread adoption.

Geng and Gao’s [31] use of an embedded fault diagnosis expert system in weapon maintenance underscores the potential of advanced AI techniques in real-time state acquisition and diagnostics. While technologically sophisticated, these methodologies necessitate substantial technical expertise and infrastructure, potentially restricting their applicability in less advanced settings.

The diversity in data types and methodologies reflects the interdisciplinary nature of research in ergonomics, design thinking, and AI/ML integration. Quantitative methods provide precision and replicability, yet they may lack depth in understanding the user experience. Mixed methods offer a balanced approach but can be resource intensive. Qualitative methods provide rich, contextual insights but are susceptible to subjective biases. Systematic reviews ensure rigorous synthesis but are dependent on existing literature. Comparative and experimental studies yield actionable insights but face challenges in generalizability. Advanced technological integrations push the boundaries of innovation but require significant resources and expertise.

3.2.2. Human Factors and Design Targets

The studies under review exhibit a profound engagement with human factors and design targets, reflecting the essential role of user-centered approaches at the intersection of ergonomics, design thinking, and AI/ML integration. A critical analysis of these elements reveals the strengths and limitations of various methodologies and their implications for practical applications.

Several studies highlight the critical role of ergonomics in improving user comfort and performance. Sung et al. [17] focused on developing predictive models for grip and key pinch strength using anthropometric data, which are essential for designing tools and equipment that reduce the risk of musculoskeletal disorders. This study directly targets ergonomic design considerations that enhance user safety and efficiency in industrial tasks by emphasizing grip strength.

Mohanty [30] also addresses ergonomic design, focusing on office chairs to reduce stress on soft tissues during prolonged sitting. By combining subjective and objective data, the study aims to create an office chair that aligns with ergonomic standards, thus improving user comfort and productivity. This research underscores the importance of ergonomic design in everyday products to enhance user well-being.

Wang and Yang [18] explored innovative design targets by evaluating motor scooter handle covers. Their approach combined user surveys with a three-phase design methodology, resulting in highly innovative concepts adaptable to market demands. This study illustrates the importance of integrating user feedback into the design process to create aesthetically pleasing and functionally superior products.

Kang [23] extended this focus on innovation by integrating neural networks to map Kansei factors to product design elements. By addressing design’s perceptual and emotional aspects, this study emphasizes the role of aesthetics and user satisfaction in product development. Using advanced engineering techniques highlights the potential for technology to enhance the design process.

Spreafico and Sutrisno [19] addressed sustainable product design using an AI-based chatbot to facilitate socially sustainable design. Their focus on the social implications of design choices highlights the importance of considering broader societal impacts in the design process. The study’s limitations, such as the chatbot’s filtering constraints, point to the need for more flexible tools that can better support designers in exploring social consequences.

Felmingham et al. [25] also emphasized human factors in the context of AI diagnostic tools for skin cancer. By integrating cognitive style, personality, and experience into the design of these tools, the study aims to enhance clinical effectiveness and user acceptance. This approach highlights the importance of considering cognitive and psychological factors in the design of AI technologies to ensure they meet users’ needs and preferences.

Mahut et al. [28] focused on interaction design by surveying interactive products. Their study established a theoretical model linking user experience, interaction, and metaphors, offering design guidelines based on human affective and cognitive processes. This research underscores the significance of intuitive user interfaces and the need for designs catering to emotional and functional user needs.

Ahmed [22] addressed smart home usability, highlighting psychological and cognitive aspects that influence user interaction with smart devices. By exploring challenges such as interoperability and user awareness, the study emphasizes the need for designs that are functional and cognitively accessible to users.

Duan, Zhang, and Chen [29] explored the application of augmented reality (AR) and Kansei engineering in product design. Their focus on enhancing perceptual and rational constraints through advanced technologies highlights the evolving nature of human factors in the context of new technological interfaces. This study points to the potential of AR to create more immersive and user-friendly design experiences.

Geng and Gao [31] addressed the maintenance of weapon equipment through an embedded fault diagnosis expert system. By focusing on the usability and efficiency of diagnostic tools, their study emphasizes the importance of designing technologies that support maintenance staff in complex and high-stakes environments.

Nandutu, Atemkeng, and Okouma [26] discussed adopting AI-based methods for mitigating wildlife–vehicle collisions, emphasizing the need to understand animal behavior to develop effective solutions. Their approach highlights the intersection of human and animal factors in design, pointing to the broader implications of behavioral insights in developing AI applications.

Whyte [27] examined the impact of AI on decision making during cyber conflicts, focusing on stakeholders’ reactions. By considering cognitive biases and the need for technical expertise, this study underscores the importance of designing AI systems that are transparent and trustworthy to users, especially in critical decision-making scenarios.

Integration of human factors and design targets across these studies highlights the multifaceted nature of ergonomic and user-centered design in AI and ML. Ergonomic considerations ensure that products are safe and comfortable for users, while design targets focus on innovation, aesthetics, and functionality. Sustainable design and social impacts are increasingly important, reflecting a broader societal awareness. Interaction design and cognitive considerations are crucial for developing intuitive and accessible user interfaces. Advanced technologies like AR and AI offer new opportunities and challenges for integrating human factors into design.

3.2.3. Main Findings and Implications

The selected papers’ main findings reveal significant insights into integrating AI/ML with design thinking and ergonomics in various domains.

Sung et al. [17] focused on developing regression and artificial neural network (ANN) models for predicting grip and key pinch strength using anthropometric data. The study’s main finding that ANN models outperform regression models in accuracy underscores the potential of advanced AI techniques in ergonomic research. This finding is significant for industrial applications, where precise ergonomic designs are crucial for reducing musculoskeletal disorders. However, the reliance on quantitative measures may overlook individual variations in comfort and usability, suggesting a need for integrating qualitative feedback to achieve a more comprehensive ergonomic assessment.

Wang and Yang [18] employed a three-phase method to generate innovative designs for motor scooter handle covers, validated through user surveys. Their finding that highly innovative concept designs are adaptable for market use highlights the effectiveness of combining user feedback with systematic design methodologies. This approach ensures that the final products are both innovative and user approved. Nonetheless, the study’s dependence on user surveys might introduce bias, as users’ self-reported preferences can sometimes be inconsistent with their actual behaviors and needs.

Spreafico and Sutrisno [19] demonstrated the capability of an AI-based chatbot to support socially sustainable design across various products. While their method proved effective, they also highlighted a significant limitation: the chatbot’s filtering mechanism, which restricted the exploration of social consequences in design. This finding underscores the need for more flexible AI tools that allow designers to investigate their work’s social impacts thoroughly. The study calls for further refinement of AI tools to better support sustainability goals in design.

Alexander and Paul [20] identified the need for improvement in digital human modeling (DHM) systems, particularly in accommodating emerging ergonomic trends and designing mobile work systems. Their systematic review reveals that current DHM systems fall short of addressing the complexities of modern ergonomic needs. This finding is critical as it points to gaps in existing technologies that must be addressed to enhance the efficacy of ergonomic interventions. Future research should focus on developing more advanced DHM systems that incorporate dynamic and context-specific ergonomic data.

Wang, Choudhary, and Yin [21] compared manual and AI-assisted design strategies in the fashion industry, finding that AI-assisted designs are preferable when innovation uncertainty is high and marginal costs are crucial. This finding emphasizes the strategic advantage of AI in managing design complexity and market variability. However, the study also suggests that manual design remains relevant in scenarios with significant fixed costs, highlighting the nuanced decision making required in adopting AI technologies. The dual approach offers a balanced perspective on the roles of AI and human designers in the creative process.

Ahmed [22] addressed the psychological and cognitive challenges of smart home integration, emphasizing the need for designs that are interoperable, reliable, and user aware. The study’s findings highlight significant barriers to user acceptance and suggest that addressing these cognitive factors is essential for successfully adopting smart home technologies. This underscores the importance of human-centered design principles in developing smart technologies, ensuring they align with users’ cognitive and psychological needs.

Kang [23] also adopted a mixed-methods approach, integrating neural networks to map key Kansei factors to product design elements. By blending qualitative and quantitative data, this study effectively captures the complexity of human factors in design, although the heavy reliance on advanced neural networks may limit the methodology’s accessibility and reproducibility for broader applications.

Mahut et al. [28] established a theoretical model linking user experience with interaction metaphors through a comprehensive survey. Their findings offer practical guidelines for designing interactive products that cater to affective and cognitive user processes. This research contributes valuable insights into how metaphors enhance user interaction and satisfaction. However, the reliance on participant feedback could introduce subjectivity, indicating a need for complementary quantitative measures to validate these findings.

Felmingham et al. [25] emphasized integrating human factors into designing AI diagnostic tools for skin cancer, focusing on potential cognitive errors and biases. Their findings underline the importance of considering cognitive styles, personality, and user experience in developing AI technologies to ensure clinical effectiveness. This critical perspective highlights the necessity of human-centered AI design, particularly in sensitive fields like healthcare, where user trust and accuracy are paramount.

Nandutu, Atemkeng, and Okouma [26] discussed the effectiveness of AI-based methods for mitigating wildlife–vehicle collisions, stressing the importance of understanding animal behavior. Their findings suggest that machine learning models can significantly enhance the detection and prevention of such collisions. This interdisciplinary approach highlights the broader applications of AI in addressing ecological and environmental challenges. However, the accuracy of these models depends heavily on the quality of behavioral data, pointing to the need for ongoing data collection and refinement.

Whyte [27] explored the impact of AI on decision making during cyber conflicts, finding that technical expertise can mitigate biases and improve evaluations of AI inputs. This study provides valuable insights into the cognitive aspects of using AI in high-stakes decision-making scenarios. The findings suggest that enhancing users’ technical understanding of AI can lead to more informed and balanced decision making. However, the experimental nature of the study may require further validation in real-world settings to confirm these results.

Duan, Zhang, and Chen [29] introduced a new system model combining augmented reality (AR) and Kansei engineering for bionic design. Their findings indicate that this integrated approach can enhance both perceptual and rational constraints in product design. This innovative method offers significant potential for creating more user-friendly and immersive design experiences. Nonetheless, the high technological demands of AR may limit its widespread adoption, suggesting a need for developing more accessible AR tools.

Geng and Gao [31] developed an embedded fault diagnosis expert system for weapon maintenance, highlighting its effectiveness in real-time state acquisition and diagnostics. Their findings suggest that such systems can significantly improve the efficiency and reliability of maintenance operations. However, the complexity and need for specialized knowledge to operate these systems may pose barriers to implementation, indicating a need for more user-friendly interfaces and training programs.

The main findings from these studies collectively underscore the diverse applications and significant potential of integrating ergonomics, design thinking, and AI/ML. While quantitative methods provide robust, objective data crucial for ergonomic design, mixed and qualitative methods offer deeper insights into user experiences and cognitive factors. The emphasis on sustainable design and social impacts reflects an evolving awareness of broader societal responsibilities. Advanced technologies like AR and AI present exciting opportunities for innovation but also highlight the need for accessibility and user-centered approaches.

3.2.4. Trends in Research

After outlining the main findings of the selected papers, Figure 5 illustrates the timeline and paper count (numbers in brackets), highlighting key points and facilitating a clearer understanding of trends through a summarized timeline. These results easily confirm the recent tendency to use AI-based models, particularly after 2022, as previously observed in the map of keywords based on average publication year (Figure 4), reflecting a growing emphasis on ethical considerations and technological advancements in design.

4. Conclusions

This research contributes to a nuanced understanding of the indispensable roles played by ergonomics and design thinking in synergy with AI/ML for product design. Ergonomics and human factors (E&HF), design thinking, AI, and ML are intricately weaving together, forming a tapestry that evolves akin to a transformative journey marked by technological (r)evolution. The findings highlight the symbiotic relationship and emphasize the crucial role of algorithmic and ergonomic integration in shaping the future landscape of generative design. Generative design is an iterative process that utilizes algorithms to generate various design solutions based on specific constraints and criteria. It uses computational models to explore numerous design possibilities, enabling designers to identify optimal solutions that balance multiple objectives, such as functionality, aesthetics, and user comfort [32]. This approach leverages advanced technologies like AI and ML to continuously refine and improve design outcomes.

Despite its potential, the concept of generative design has not been extensively covered in the reviewed literature, highlighting an area for future research to explore its implications and applications further. By integrating ergonomic principles into generative design, the potential for creating highly functional and user-friendly designs is significantly enhanced, marking a pivotal shift in design methodologies.

The study highlights these disciplines’ resilience and lasting relevance in the ever-evolving design field, despite the potential limitations associated with contemporary web crawlers. The selected studies’ diverse approaches underscore the need for a multifaceted understanding of how these fields interact to create innovative, user-centered products, providing a stable foundation in the midst of constant change.

Several limitations were identified in this research. The reliance on contemporary web crawlers and the varying quality of the available literature may have impacted the comprehensiveness of the review. Additionally, the studies reviewed often varied in their methodological rigor and scope, which may affect the generalizability of the findings. Notably, none of the reviewed papers addressed the secular trend of population weight growth, which is crucial for everyday product design ergonomics. Future research should address these limitations by employing more robust search methodologies and expanding the scope of the review to include a wider range of studies, particularly those focusing on population trends and their implications for ergonomics.

The diversity in data types and methodologies reflects the interdisciplinary nature of ergonomics, design thinking, and AI/ML integration research. This critical analysis highlights the strengths and limitations of each approach, emphasizing the need for methodological pluralism to capture the complex interplay between human factors, design, and emerging technologies. Quantitative, qualitative, and mixed methods each contribute unique insights, reinforcing the value of integrating multiple perspectives to enhance the robustness of research outcomes.

The critical analysis of human factors and design targets underscores the importance of a holistic approach to design. Incorporating ergonomic principles, user feedback, and advanced technological capabilities is essential for creating innovative and user-friendly products. This comprehensive approach ensures that designs meet aesthetic and functional requirements and enhance user comfort, safety, and overall experience. The findings stress the importance of addressing both physical and cognitive aspects of user interactions to achieve truly effective design solutions.

The main findings from these studies collectively underscore the diverse applications and significant potential of integrating ergonomics, design thinking, and AI/ML. While quantitative methods provide robust, objective data crucial for ergonomic design, mixed and qualitative methods offer deeper insights into user experiences and cognitive factors. The emphasis on sustainable design and social impacts reflects an evolving awareness of broader societal responsibilities. Advanced technologies like AR and AI present exciting opportunities for innovation but also highlight the need for accessibility and user-centered approaches.

Future research should continue exploring the intersections between ergonomics, design thinking, and AI/ML, focusing on developing more sophisticated models and tools to optimize the design process further and enhance user satisfaction. There is a need for longitudinal studies that examine the long-term impacts of these integrated approaches on product design and user experience. Additionally, future studies should investigate the ethical implications of AI/ML in design, ensuring that these technologies are used responsibly and sustainably. Moreover, research should expand to include the secular trend of current population weight growth, as this is an important area for the further direction of ergonomics in everyday product design.

Overall, this research demonstrates the transformative potential of combining ergonomics, design thinking, and AI/ML in product design. By leveraging the strengths of each discipline, more adaptive, efficient, and user-centric products can be created. The insights gained from this study provide a valuable roadmap for future research and development in this interdisciplinary field.