Constructing a Sustainable Evaluation Framework for AIGC Technology in Yixing Zisha Pottery: Balancing Heritage Preservation and Innovation

Abstract

This study develops a sustainable evaluation framework for Yixing Zisha pottery design schemes generated by Artificial Intelligence Generated Content (AIGC) technology, emphasizing the integration of cultural heritage preservation with innovation. As a traditional Chinese craft and a recognized element of intangible cultural heritage (ICH), Yixing Zisha pottery is celebrated for its cultural depth and unique design techniques. Guided by emotional design theory, the framework assesses aesthetic, functional, and emotional dimensions through hierarchical analysis. Using the Delphi method and Analytic Hierarchy Process (AHP), primary and secondary indicators were identified and weighted based on expert consensus. AIGC technology, underpinned by advanced AI algorithms, generates culturally authentic yet innovative design solutions, striking a balance between tradition and modernity. The findings reveal that this approach enhances design diversity, functionality, and efficiency while fostering global cultural awareness. By providing practical guidance for integrating AIGC technology into traditional craftsmanship, the research offers a replicable model for other traditional crafts and contributes to the theoretical advancement of sustainable cultural heritage practices. By bridging the gap between digital innovation and heritage preservation, this study addresses the critical need for sustainable strategies in the creative industries.

1. Introduction

Yixing Zisha pottery, a prominent representative of China’s intangible cultural heritage, epitomizes the culmination of traditional craftsmanship, with its profound artistic, cultural, and functional value. Renowned for its unique clay materials, refined design styles, and unparalleled artisanal techniques, Yixing Zisha pottery has gained global recognition as an emblem of Chinese traditional culture. However, as globalization accelerates, the industry faces challenges such as design homogenization, declining market competitiveness, and an inability to adapt to modern consumer preferences [1]. These challenges pose critical threats to the sustainability of Yixing Zisha as a cultural heritage. Balancing the preservation of its cultural significance with the integration of modern innovation is not only a pressing issue for this craft but also for other forms of intangible cultural heritage facing similar pressures [2].

In recent years, AIGC technology has emerged as a transformative tool in the creative industries [3]. AIGC refers to the use of artificial intelligence to generate and create content, aiming to produce high-quality outputs based on user-inputted keywords or requirements, partially replacing or assisting human creation [4]. This technology relies on large-scale pre-trained models and is applied across domains including text, image, video, and other multimodal content creation, demonstrating strong potential in the digital economy, media production, and cultural communication [5]. Leveraging deep learning and big data analytics, AIGC enables the automated generation of design schemes, offering both creativity and efficiency in fields such as visual arts, advertising, and industrial design [6]. However, when applied to Yixing Zisha pottery—where cultural richness, artisanal uniqueness, and symbolic meaning are deeply embedded—AIGC’s applicability faces significant challenges. Figure 1 showcases a Yixing Zisha teapot design generated by the AIGC software DALL E 3, utilizing the theme of a dragon motif. The design exhibits remarkable attention to detail in both its form and texture, closely resembling a real Yixing Zisha teapot. The intricate dragon relief and precise modeling on the surface highlight the software’s capacity to integrate cultural elements with aesthetic sophistication. Additionally, the material simulation effectively mimics the characteristic clay texture unique to Yixing Zisha pottery, reinforcing the authenticity of the design.

This design reflects the transformative potential of AIGC technology in traditional crafts, particularly in automating the creation of culturally rich artifacts. Traditional craftsmanship such as Yixing Zisha is not merely about the creation of physical artifacts but also embodies profound cultural narratives and emotional resonance. Feedback from practitioners and experts reveals a shared concern that AIGC technology must respect and faithfully convey the cultural essence and symbolic meanings intrinsic to Yixing Zisha pottery. This highlights a dual challenge: advancing the creative potential of AIGC while preserving the intangible heritage embedded in traditional crafts [7]. In response, the present study incorporates Emotional Design Theory as a theoretical lens to tackle this challenge. Emotional Design Theory evaluates designs across three dimensions: aesthetics, functionality, and reflectiveness. In this study, it serves as the theoretical foundation for the evaluation framework, guiding the assessment of AIGC-generated designs [8]. To adapt Emotional Design Theory for AIGC’s iterative and generative processes, this study introduces an additional dimension—“improvement potential and flexibility”—to capture how AI-driven tools refine and optimize designs over time. This expanded framework aligns traditional craft values with modern technological capabilities. This study expands Emotional Design Theory beyond conventional design attributes, aligning it with the multidimensional requirements of blending innovation and heritage [9].

Despite the conceptual advantages offered by Emotional Design Theory, current evaluation standards for Yixing Zisha remain rooted in traditional craftsmanship criteria and lack a systematic approach that validates modern technological applications. This gap restricts scientific assessment of how AIGC-generated designs convey Yixing Zisha’s cultural essence and symbolic richness, thereby limiting their potential to rejuvenate the craft in contemporary contexts. Hence, the study develops a sustainable evaluation framework that integrates Emotional Design Theory, the Delphi method, and AHP to systematically bridge technological innovation and cultural heritage preservation. Critically, this approach provides a replicable model for other traditional crafts undergoing digital transformation. The Delphi method, recognized for its iterative approach in synthesizing expert consensus, enables the identification of key evaluation criteria for AIGC-generated designs [10]. These criteria are then integrated into the extended theoretical framework guided by Emotional Design Theory. AHP complements this by rank-ordering those criteria, offering a quantitative structure for comprehensive and structured decision-making [11]. Together, these methods reinforce the framework’s ability to evaluate both traditional and innovative design dimensions, ensuring its applicability to diverse contexts. Together, these methods form a multidimensional evaluation system that captures the interplay between cultural significance and technological capabilities. We propose an integrative model (later discussed and illustrated) that combines Emotional Design Theory with an additional layer reflecting AIGC’s iterative flexibility. This extended theoretical framework not only highlights the aesthetic, functional, and reflective dimensions but also incorporates “improvement potential and flexibility”, thereby ensuring that Yixing Zisha designs remain culturally authentic while embracing modern innovation. Hence, the present study advances both theory and practice by offering a systematic framework—rooted in expert feedback and quantitative weighting—to scientifically assess the applicability of AIGC in Yixing Zisha pottery design. By integrating Emotional Design Theory with iterative and quantitative methods, the proposed framework ensures cultural authenticity while fostering sustainable innovation. This framework is not only relevant for Yixing Zisha but also adaptable to other traditional crafts undergoing digital transformation. Ultimately, the proposed system provides a scientific basis for applying AIGC technology in Yixing Zisha design and offers practical references for modernizing traditional crafts. Ultimately, our objective is to foster sustainable innovation within traditional craftsmanship while safeguarding its cultural heritage, aligning with broader goals of heritage conservation and sustainable development [12].

2. Literature Review

Artificial Intelligence Generated Content technology, with its advanced capabilities in deep learning and big data processing, has revolutionized the creative industries [13]. AIGC significantly enhances design efficiency by automatically generating design schemes, driving diversity and innovation in visual creativity, and becoming an indispensable tool in modern design processes [14,15]. Gao et al. (2022) pointed out that the widespread application of AIGC has significantly improved creative production efficiency, especially in fields requiring rapid iteration and mass production [16]. Li and Zhang (2024) proposed that AIGC, after specialized training, can generate virtual humans within minutes or even seconds. These virtual humans, leveraging their understanding of designs and works, assist designers in completing their design tasks [17]. With advancements in manufacturing technologies, functional differences among various products have gradually diminished. However, as consumer demands diversify, designing product forms that meet consumers’ emotional needs has become increasingly critical [18]. In a “consumer-oriented” market environment, users are gradually placing more emphasis on the aesthetic value and cultural connotations embodied in a product’s appearance, in addition to its practical functions [19]. Meanwhile, Huang and Zheng (2022) emphasized that integrating AIGC with human–computer collaboration systems not only improves the quality of creativity but also significantly reduces designers’ workload, allowing them to focus more on concept development and design optimization [19]. These studies indicate that AIGC’s technical features are redefining traditional creative production models and elevating its importance in commercial design. Further exploring design practices, scholars have highlighted notable differences between design structure and design patterns in application. Hassan et al. (2021) noted that the application of design pattern collections provides designers with reliable tools, particularly in balancing aesthetic and performance demands [20]. This perspective underscores that AIGC is not merely a tool for generating creativity but also a structured solution for addressing constraints between creative expression and technical requirements.

Although AIGC technology has matured in the field of commercial design, its application in cultural heritage preservation and traditional craft design remains in the exploratory stage. Cultural heritage design is far more complex than commercial design, as it requires not only aesthetic appeal but also the preservation of cultural symbols’ integrity and depth of artistic expression while maintaining a close connection with social and historical contexts [21,22]. Existing research shows that AIGC demonstrates immense potential in cultural heritage preservation and offers new approaches for integrating traditional crafts with modern technology. For example, in the restoration of the Bishutang Hall scene at the Chengde Mountain Resort, Yi et al. (2024) utilized AIGC combined with digital modeling to achieve a highly accurate restoration of visual, stylistic, and historical consistency, providing critical technological support for cultural heritage preservation [23]. The Dunhuang Academy utilizes AIGC technology to assist in the construction of digital scenes by performing image recognition and semantic analysis on massive amounts of digital image data, enabling virtual scene exploration for visitors. Through AIGC technology, visitors can observe the uniqueness and rich content of each cave from any angle, providing an immersive travel experience [24]. Additionally, AIGC has shown remarkable results in promoting innovation in traditional crafts. For instance, He and Zhao (2024) demonstrated how AIGC generates diverse design schemes, driving the organic integration of traditional craftsmanship and modern design elements in Suzhou jade carving [25]. Yang and Chen (2024) examined AIGC’s innovative applications in color and style modernization, using Taohuawu woodblock prints as a case study [26]. Wang et al. (2024) developed a specialized database for Ming-style furniture using the Midjourney platform. Their empirical research validated and reproduced the aesthetics of Ming-style furniture, demonstrating its cultural significance. By leveraging AIGC technology, they advanced both traditional and contemporary furniture design, providing sustainable tools for industry development and cultural heritage preservation [27]. Their findings provide insights into the fusion of intangible cultural heritage with contemporary design practices. In today’s resurgence of traditional culture, the integration of AIGC, an emerging technology, with traditional industries will profoundly impact the preservation of traditional craftsmanship and innovation in the cultural heritage field. The fusion of artificial intelligence with traditional culture will inevitably expand and elevate the pathways for the inheritance and innovation of traditional craftsmanship, fostering new avenues for its development [28]. Anwar R and Mohamed Raif D (2022) highlighted that in cultural heritage design, users, as aesthetic receivers, often prioritize the importance of a product’s performance over its mere visual appeal [29]. This emphasis on balancing functionality and aesthetics offers critical insights for AIGC’s further development and provides a new direction for its practical application in cultural heritage design. As a highly intelligent design assistance technology, AIGC continuously performs detailed evaluations and judgments during the design output process. Through iterative refinement and feedback on various outcomes, it precisely determines and optimizes the final solution [30].

Despite its significant potential, AIGC technology faces challenges in cultural heritage design. Fan and Sun (2021) pointed out that AIGC often distorts cultural symbols due to limitations in semantic understanding when generating design content [31]. Its inability to accurately grasp the semantic depth of symbols can lead to superficial representations, potentially weakening the cultural significance of the design and even misleading audiences’ perceptions of traditional culture. Zong et al. (2024) further emphasized that AIGC-generated content often lacks cultural depth due to its dependence on existing datasets, simplifying complex cultural elements into single visual symbols and failing to reflect the spiritual value and cultural richness of traditional crafts [32]. Chen et al. (2023) argue that AIGC technology can, to some extent, achieve an organic integration of modern technology and traditional art. However, relying entirely on machine systems for creation may lack the personal style and emotional expression of artists. Since AIGC technology originates from Western science and technology, developing indigenous generative AI platforms raises critical questions about how to adapt and respect the profound cultural context. Particularly in the domain of traditional art and craftsmanship, it is essential to explore the boundaries of platform applicability to ensure it neither infringes upon nor dilutes the uniqueness of traditional culture [33]. These limitations are particularly pronounced in the field of intangible cultural heritage design, which places a greater emphasis on emotional resonance and cultural identity.

In the face of advancements and disruptions brought by technology, Yixing Zisha art is actively seeking to rejuvenate itself in a dynamic and innovative manner [34]. Yixing Zisha ceramics, as a prominent representative of China’s intangible cultural heritage, showcases exquisite craftsmanship and profound cultural connotations. The design of Zisha ceramics demands not only precise craftsmanship but also the conveyance of deep cultural narratives and symbolic significance. However, studies by Chen et al. (2023) and Li et al. (2023) revealed that while AIGC excels in generating intricate patterns, it falls short in reproducing the unique details and textures inherent to Zisha ceramics [35,36]. This technical limitation not only restricts the artistic expression of the generated content but also exacerbates homogenization, undermining design innovation. Liang (2023) further pointed out that AIGC’s overreliance on existing datasets confines its outputs to traditional stylistic frameworks, making it difficult to break design boundaries [37]. Tang (2023) pointed out that AIGC technology still relies on the wisdom of human artists to build extensive databases and uses complex computational power to simulate generation. Regardless of how it evolves, countless traces of human artists’ influence can still be seen in its creations [38]. The relationship between AIGC technology and Yixing Zisha pottery, particularly in navigating the balance between tradition and innovation, represents a critical focal point of this study.

Emotional Design Theory offers an important theoretical framework for addressing these challenges. As illustrated in Figure 2, Norman’s three levels of emotional design, proposed by Don Norman, categorize user experience into instinctive, behavioral, and reflective levels, each encompassing specific design attributes and considerations relevant to this study [39]. The instinctive level focuses on immediate sensory and aesthetic appeal, including elements such as first impressions, visual harmony, material textures, color matching, and craftsmanship precision [40]. This is particularly relevant to the design of Yixing Zisha ceramics, which are renowned for their exquisite details and elegant forms. These elements play a crucial role in engaging users on an emotional and sensory level, establishing a direct connection with the craftsmanship and artistry of the pottery. The behavioral level emphasizes functionality and usability. It addresses practical aspects such as user interaction, ergonomics, and practicality [41]. These factors are essential for ensuring that Yixing Zisha ceramics maintain their traditional craftsmanship while adapting to the functional needs of modern users. The reflective level delves into deeper cultural and emotional dimensions, encompassing cultural identity, emotional resonance, symbolic meaning, narrative depth, and cultural expression. In the case of Yixing Zisha ceramics, this level is critical as it highlights the cultural narratives, traditional artistry, and emotional depth that make these ceramics an enduring representation of intangible cultural heritage. By aligning with users’ values and cultural expectations, designs at this level ensure the preservation and continuation of the profound cultural significance embedded within Yixing Zisha.

Wang and Zhang (2023) proposed that in the current social context, in addition to aspects such as form, patterns, and craftsmanship, the cultural and artistic dimensions of ceramic products still require further exploration. By deeply studying and addressing users’ emotional needs, the design concepts of ceramic culture can be broadened, cultural confidence enhanced, and ceramic cultural and creative designs made to better meet the emotional needs of their audience [42]. This framework provides a foundational approach to evaluating designs, balancing aesthetic, functional, and cultural dimensions. In cultural heritage design, the reflective level is particularly critical because it focuses on cultural resonance and emotional identification [43]. Studies by Lin et al. (2013) and Chen and Chen (2020) demonstrated that the reflective level not only satisfies users’ cultural identity needs but also significantly enhances the emotional depth of cultural product design [44,45]. In AIGC-generated Yixing Zisha ceramic designs, Emotional Design Theory provides a multi-dimensional evaluation framework for assessing aesthetic appeal, functionality, and cultural expression. Tang, Zhang, and Hu (2023) conducted a study on Yixing Zisha pottery, introducing cultural elements into practice. They extracted the contours of cultural shapes and established the relationship between contour modeling and element models. Guided by aesthetic demands, they optimized the design process to achieve the best design solutions [46]. Its incorporation ensures that AIGC-generated content excels in artistic innovation while faithfully conveying the cultural connotations and symbolic significance of Zisha ceramics. To scientifically evaluate the application of AIGC technology in Yixing Zisha ceramic design, this study integrates the Delphi method and the AHP method to construct an evaluation system. In social science research, the Delphi method and the AHP method are two commonly used methods, employed for aggregating expert opinions and conducting multi-criteria decision analysis, respectively [47,48]. When these two methods are combined, they can effectively enhance the quality and reliability of decision-making. Such a combination has proven valuable in designing evaluation systems across various fields [49]. For instance, Taleai and Mansourian (2008) used the Delphi method to collect expert opinions and applied AHP to assign weights, thereby scientifically pinpointing issues such as policy, funding, and public participation in urban planning [50]. Similarly, Lin (2011) employed Delphi to identify design requirements and used AHP to assign weights to indicators like cultural symbolism and comfort in a study on Hakka cultural creative industries in Taiwan [51]. These cases strongly support scientific decision-making in cultural heritage design. Building on these successful precedents, the present study utilizes Delphi to gather expert opinions—ensuring the scientific nature and applicability of the evaluation indicators—and employs AHP for weight assignment, making the entire evaluation system both structured and quantitative. Ultimately, the proposed system provides a scientific basis for applying AIGC technology in Yixing Zisha design and offers practical references for modernizing traditional crafts.

In developing this integrated framework (Figure 3), the study draws on Norman’s Emotional Design Theory, emphasizing aesthetic, functional, and cultural dimensions while addressing the iterative capabilities unique to AIGC [52]. This theoretical foundation provides a robust structure for analyzing how AIGC technology influences traditional crafts like Yixing Zisha pottery, allowing for a multi-dimensional evaluation across instinctive, behavioral, reflective, and extended layers. Historically, evaluations of Yixing Zisha pottery have predominantly focused on aspects such as form, compositional characterization, and craftsmanship [53,54,55]. However, with the adoption of AIGC technology as a design aid, such an approach is evidently insufficient to address the broader complexities of integrating emerging technologies with traditional crafts [56]. In the context of AIGC technology, it is imperative to establish a multidimensional framework that facilitates design innovation while balancing the dual objectives of leveraging advanced technological capabilities and preserving the cultural and artistic essence of Yixing Zisha pottery. The addition of an extra layer—improvement potential and flexibility—was inspired by both practitioner feedback and AIGC’s technological features. This layer reflects the continuous refinement, adaptability, and post-modification enhancements that AI-based design enables. Practitioner feedback during the preliminary investigation highlighted a strong demand for iterative and adaptive features, as many industry professionals emphasized AIGC’s capacity to evolve and optimize designs over time to meet changing market and user needs. This practical input supports the theoretical expansion of Emotional Design Theory, ensuring that the framework comprehensively addresses both static and dynamic aspects of design evaluation. By uniting these four levels within a single framework, the study offers a comprehensive tool for evaluating a craft’s immediate aesthetic and functional attributes alongside its deeper cultural values and capacity for adaptation over time. The framework bridges the gap between the fixed nature of traditional evaluation frameworks and the dynamic, generative nature of AIGC processes. Moreover, it aligns with the existing literature, which emphasizes the dual imperatives of creative efficiency and cultural preservation in the application of AIGC technologies to heritage crafts [28,57]. This extended framework provides a valuable lens for examining the interplay between cultural heritage and technological innovation, highlighting how AIGC can sustain and innovate within traditional crafts like Yixing Zisha pottery. It not only ensures a balance between tradition and modernity but also offers a theoretical contribution by adapting Emotional Design Theory to the era of AI-driven design.

3. Research Methodology

3.1. Research Design

This study aims to develop a multidimensional evaluation system for assessing the performance of AIGC technology in Yixing Zisha pottery design. By drawing upon Emotional Design Theory and incorporating practitioners’ practical feedback, the research design employs the Delphi method and AHP to scientifically screen and assign weights to evaluation indicators. This approach ensures scientific rigor, comprehensiveness, and practical relevance. In particular, the Delphi method and AHP were selected for their complementary strengths: the Delphi method facilitates structured and iterative refinement of evaluation indicators based on expert consensus, while AHP quantitatively prioritizes these indicators, ensuring objectivity in weight distribution. Such integration is well-suited to AIGC-driven heritage design, where cultural authenticity and technological capabilities must both be considered.

This research also builds on the theoretical framework proposed in the literature review, which integrates Emotional Design Theory’s three core dimensions—instinctive, behavioral, and reflective layers—with an additional layer, improvement potential and flexibility. This theoretical structure guides the design and refinement of evaluation indicators, ensuring the framework captures both the cultural essence and adaptive capabilities introduced by AIGC technology.

To ensure the comprehensiveness of the evaluation system, a questionnaire survey was conducted to gather practitioners’ concerns regarding AIGC technology in Yixing Zisha design. These findings, reflecting traditional craft values and modern digital demands, were instrumental in constructing the preliminary indicators. Specifically, practitioners’ key focus areas—visual appeal, cultural heritage expression, functional innovation, and improvement potential—were identified through text analysis and combined with the core dimensions of Emotional Design Theory. Hence, the study emphasizes adaptability and flexibility, recognizing that AIGC introduces iterative elements that traditional frameworks may overlook.

As shown in Figure 4, the overall research design consists of three key phases:

1. Definition of Evaluation Dimensions: Based on the survey results and inspired by Emotional Design Theory, the evaluation system was initially categorized into instinctive, behavioral, and reflective dimensions. Aspects not fully addressed by Emotional Design Theory, such as adaptability and flexibility, were introduced under “improvement potential and flexibility” to capture AIGC’s iterative nature.

2. Indicator Screening and Refinement: The Delphi method was used to refine the preliminary indicators through multiple rounds of expert feedback, achieving a high degree of consensus while addressing ambiguities and redundancies.

3. Indicator Weight Assignment: AHP was employed to assign weights to each indicator within the hierarchical model. Pairwise comparisons and consistency tests were conducted to ensure reliability in the weight assignments.

This research establishes a robust evaluation framework by combining practitioners’ feedback with Emotional Design Theory and refining the indicators using the Delphi method and AHP. Such a dual-method approach systematically balances cultural preservation with technological innovation, aligning the study with its objective of bridging heritage and modernity.

3.2. Indicator Development

3.2.1. Preliminary Indicator Identification

Estimating preliminary indicators forms a crucial foundation for constructing the evaluation system. This study adopted a systematic approach to ensure scientific and practical relevance, combining Emotional Design Theory with feedback from Yixing Zisha pottery practitioners. Although Emotional Design Theory typically highlights sensory appeal, functional performance, and cultural resonance in design works, it has proven effective in evaluating cultural adaptability and emotional depth in cultural heritage design [58]. In this context, the study also considered feedback emphasizing adaptability and iterative improvement, aligning with the capabilities of AIGC. To obtain a broad range of perspectives, data collection involved online and offline methods—online surveys via the Wenjuanxing platform and offline surveys at local ceramic markets, craft exhibitions, and specialized colleges in Yixing. This approach ensured broad participation and minimized bias. In total, 315 valid responses were received, revealing that practitioners’ concerns centered on visual performance, functional optimization, cultural heritage expression, and design flexibility. Stratified random sampling guaranteed proportional representation across age, professional experience, and role type. By systematically analyzing the questionnaire responses, the researchers identified high-frequency themes, including visual attractiveness, color coordination, material texture, functionality, user interaction, cultural identity, heritage value, technological feasibility, and flexibility.

Aligned with the typical three-level structure of Emotional Design Theory, this study categorized the insights into four primary dimensions. First, the instinctive level focuses on sensory appeal, visual effects, and aesthetic performance. Second, the behavioral level emphasizes functionality and usability. Third, the reflective level deals with cultural resonance and emotional expression. Additionally, a fourth dimension, “improvement potential and flexibility”, was introduced to address the iterative and generative aspects associated with AIGC technology. These inclusions—not arbitrarily chosen—were derived from practitioner surveys pointing to adaptability (modifiability), feasibility, and enhancement potential as critical factors. This aligns with AIGC’s unique ability to support iterative design improvements, distinguishing it from static design evaluation frameworks. Consequently, the classification process connected theory and practice, ensuring that the preliminary indicator system addressed the multi-level demands of AIGC in Yixing Zisha design and laid a solid foundation for subsequent optimization via the Delphi method.

3.2.2. Refinement Through Delphi Method

Building upon the preliminary indicator extraction, this study employed the Delphi method to refine and optimize the evaluation indicator framework systematically. Chosen for its ability to gather and synthesize expert opinions iteratively, the Delphi method ensures that evaluation indicators align with the diverse demands of AIGC-generated Yixing Zisha designs. Widely used for constructing multidimensional evaluation systems, Delphi relies on multiple rounds of anonymous feedback to achieve consensus. This iterative process yielded a robust set of core evaluation indicators. As shown in

Table 1. Basic information on experts.

Project	Frequency	Percentage %	Project	Frequency	Percentage %
Gender			Areas of Expertise
Male	10	62.5	AIGC Technology	4	25
Female	6	37.5	Yixing Zisha Design/Production	12	75
Age			Years of working experience
30–39	4	25	10–14 years	3	18.75
40–49	4	25	15–19 years	3	18.75
50–59	4	25	≥20 years	10	62.5
60–69	4	25	Professional ranks
Academic qualifications			Intermediate title	4	25
Undergraduate degree	9	56.25	Deputy senior title	5	31.25
Master’s degree	3	18.75	Senior title	7	43.75
Doctorate	4	25

, 16 experts were invited to participate, with the number of participants deliberately selected to balance diversity and manageability, aligning with Dalkey and Helmer’s (1963) recommendation that 10 to 20 participants form an effective Delphi panel [59]. Experts were recruited via professional networks, industry recommendations, and academic collaborations, and their credentials were verified based on qualifications and publications. This strategy maintains breadth of perspectives while ensuring accuracy in feedback analysis.

Additionally, the panel composition demonstrated significant diversity in gender, age, professional experience, academic qualifications, and professional ranks. Their professional ranks ranged from intermediate to senior, with 31.25% holding deputy senior titles. Experts were selected from two primary areas of expertise: AIGC technology (25%) and Yixing Zisha design/production (75%). Their working experience highlighted substantial depth, with 62.5% having over 20 years of experience. This diversity provided comprehensive insights into the functional, technical, cultural, and artistic dimensions of AIGC technology in Yixing Zisha design. This representative composition ensured that the Delphi method achieved a balance between practical expertise and academic rigor, reinforcing the relevance of the evaluation indicators to the multifaceted demands of integrating AIGC technology with traditional craftsmanship.

To enhance the effectiveness of the Delphi process, the research team distributed questionnaires via email and instant messaging tools, conducting two rounds of expert consultations. In the first round, experts rated the preliminary indicators based on aesthetic value, functional requirements, and cultural expression while offering suggestions for modification or addition. The research team systematically analyzed the feedback, calculating the mean and coefficient of variation (CV) for importance ratings and assessing the consensus using Kendall’s W coefficient. The initial results indicated that while most indicators aligned well with the application of AIGC technology in Yixing Zisha design, some required further refinement in phrasing or categorization.

Following the first round, the research team revised the preliminary indicators, merging overlapping items and adding dimensions like “creative forms” to better capture the demand for innovative designs. In the second round, the revised indicators were resubmitted for expert evaluation. The results demonstrated improved mean scores, significantly reduced CVs, and enhanced consistency among expert opinions. The final indicator framework comprised 4 primary dimensions and 17 secondary indicators, comprehensively addressing the multidimensional applications of AIGC technology in Yixing Zisha design. The iterative feedback process, validated through metrics such as Kendall’s W coefficient, demonstrated an enhanced consensus, culminating in a refined set of indicators that met both theoretical and practical requirements. The application of the Delphi method in this study not only enhanced the scientific rigor and applicability of the indicators but also ensured objectivity and consensus through its iterative feedback mechanism. The finalized indicator framework provides a solid theoretical foundation and practical guidance for subsequent quantitative analysis and model validation.

3.2.3. Weight Allocation with AHP

The allocation of indicator weights is a critical step in evaluating AIGC-generated Yixing Zisha teapot design schemes. A scientifically and reasonably distributed weighting ensures a balanced assessment across multiple dimensions, highlighting both cultural heritage and technological innovation. The application of AHP in this study provided a structured, quantitative approach to allocate weights to evaluation indicators identified through the Delphi process. This integration ensured the scientific reliability and robustness of the evaluation framework.

As shown in Figure 5, a hierarchical model consisting of three levels—goal layer, criterion layer, and scheme layer—was constructed during the weight allocation process. Each of these layers fulfil a unique role within the AHP process, contributing to a comprehensive and methodical approach to addressing complex decision-making problems. The goal layer represents the overall objective of the decision, guiding the entire analytical process. The criterion layer outlines the various factors or criteria that influence the decision, enabling a detailed evaluation of each aspect’s relative importance. Finally, the alternative layer encompasses the possible options or solutions under consideration, allowing for a systematic comparison based on the criteria established. By segmenting the model into these distinct layers, AHP ensures a structured and logical allocation of weights. This hierarchy not only clarifies the relationships between different decision elements but also provides the flexibility needed to integrate both cultural and technological aspects, as guided by Emotional Design Theory. This hierarchical structure, guided by the three-layer framework of emotional design theory, integrates cultural and technological aspects, providing theoretical and practical support for comprehensively evaluating AIGC-generated designs.

Based on the hierarchical structure, pairwise comparisons were conducted through judgment matrices to determine the relative importance of each indicator. As shown in

Table 2. Saaty’s 1–9 scale method.

Scale Value	Definition	Explanation
1	Equal Importance	Both factors contribute equally to the objective.
3	Slightly More Important	Based on experience and judgment, one factor is slightly more important than the other.
5	Clearly More Important	Based on experience and judgment, one factor is clearly more important than the other.
7	Strongly More Important	One factor is strongly more important than the other, and this importance is significantly evident.
9	Extremely More Important	One factor is extremely more important than the other, reflecting the greatest degree of superiority.
2, 4, 6, 8	Intermediate Values	Used for judgments that fall between the standard values, indicating that the importance of the two factors lies between the defined levels above.
1/n	Reciprocal Value	If one factor is judged to be of importance n compared to another, the reciprocal value of the comparison is 1/n.

, Saaty’s 1–9 scale was used as a standardized tool to facilitate these comparisons [60]. This scale allows experts to express the relative importance of two indicators in a consistent and structured manner [61]. Each numerical value corresponds to a qualitative judgment, ranging from “equal importance” (1) to “extremely more important” (9), with intermediate values (2, 4, 6, 8) enabling finer distinctions in evaluation. Reciprocal values (e.g., 1/3, 1/5) are applied when one factor is deemed less important than another, ensuring logical symmetry in the pairwise comparison process. Experts evaluated the indicators in the criterion and scheme layers, generating judgment matrices based on their perceived significance in AIGC design applications. This process ensured that the subjective judgments of experts were systematically quantified, forming the foundation for calculating indicator weights. The structured approach of the 1–9 scale minimizes biases and enhances the consistency of evaluations.

After constructing the judgment matrices, the eigenvector method was employed to calculate the relative weights of the indicators. This mathematical method derives the priority vector (weights) by solving the principal eigenvector of the matrix, ensuring that the results accurately reflect the relative importance assigned by experts. Consistency checks were performed to ensure logical consistency in expert judgments. The consistency index (CI) and the consistency ratio (CR) were calculated following Saaty’s guidelines to validate the matrices. Specifically, a CR value below 0.1 indicates acceptable consistency, confirming the reliability of the judgments. If this threshold is exceeded, adjustments to the matrix are made, and the weights are recalculated to ensure logical coherence and accuracy. This systematic process ensures that the evaluation framework is both scientifically robust and adaptable to the specific demands of AIGC technology in Yixing Zisha design. For instance, in determining primary indicator weights, the instinctive level might be assigned greater weight due to its emphasis on sensory appeal, while the behavioral and reflective levels provide complementary insights into functionality and cultural value.

The finalized weights are applied in the subsequent scoring of the design schemes. The scoring process typically uses a Likert five-point scale, where experts evaluate each design scheme’s performance across different dimensions. Each score is multiplied by its corresponding weight to calculate the overall performance score of the design scheme. This process ensures that the final assessment reflects the multidimensional priorities of the study, combining cultural heritage preservation with technological innovation.

This quantitative integration ensures that assessments are replicable, objective, and aligned with the multidimensional objectives of the study. By employing Saaty’s 1–9 scale and integrating it with rigorous consistency checks and eigenvector calculations, this study provides a robust and scientifically validated foundation for evaluating AIGC-generated design schemes in the context of Yixing Zisha pottery.

4. Data Analysis and Results

4.1. Initial Insights from Practitioners’ Feedback: Foundation for Developing Evaluation Indicators

In the initial stage of constructing the evaluation indicator system, this study conducted an open-ended questionnaire to extensively collect feedback from Yixing Zisha pottery practitioners regarding their key concerns on AIGC technology in design. The target respondents included a diverse group of practitioners—experienced artisans, emerging designers, mid-level professionals, and apprentices—ensuring a representative sample that reflects both traditional and innovative perspectives within the industry. A detailed word frequency analysis was then performed on the responses, and the findings were systematically aligned with the framework of the multidimensional evaluation framework, which extends Emotional Design Theory. The analysis revealed that practitioners’ core concerns closely relate to the instinctive, behavioral, and reflective layers of design experience, providing theoretical and data-driven support for developing preliminary evaluation indicators.

In total, 315 valid responses were collected, encompassing a wide range of expertise levels—from experienced artisans to mid-level professionals—further ensuring the sample’s representativeness. Data collection involved online and offline methods for inclusivity: offline distribution targeted artisans less familiar with digital platforms, addressing potential participation barriers. The questionnaire aimed to identify the key focus areas and practitioners’ expectations regarding AIGC in Zisha design, posing core questions about appearance, material, practicality, feasibility, and design suggestions. The core questions were as follows:

1. In applying AIGC technology to Zisha design, which aspects would you pay more attention to (e.g., appearance, material, etc.)? Please describe in as much detail as possible.

2. Regarding the application of AIGC technology in Zisha design, which performance aspects do you focus on (e.g., practicality, feasibility, etc.)? Please describe in as much detail as possible.

3. What suggestions do you have for the application of AIGC technology in Zisha design? Please share your insights and suggestions.

The research team combined manual coding with keyword frequency analysis to ensure robust analysis. Two researchers independently reviewed each response to validate the identified themes, enhancing the reliability of the results. Through systematic classification, the study identified key focus areas expressed by practitioners and organized them within the expanded Emotional Design Theory framework. As shown in

Table 3. List of the top 50 high-frequency keywords.

No.	Word	Frequency	No.	Word	Frequency
1	Culture	136	26	Functionality	14
2	Design	123	27	Consummate	14
3	Technology	93	28	Intelligence	13
4	Zisha	92	29	Designer	13
5	Appearance	65	30	Improvement	12
6	Accuracy	64	31	Material Texture	12
7	Potential for Improvement	41	32	Field	12
8	Value	34	33	Recognition	11
9	Feasibility	30	34	Integration	11
10	Cultural Heritage	29	35	Texture Feel	11
11	Utilization	28	36	Element	11
12	Promotion	26	37	Unique	11
13	Aesthetic	26	38	Aesthetic Design	11
14	Modeling	25	39	Detail	10
15	Practicality	24	40	Harmony	10
16	Tradition	24	41	Process	10
17	Color Matching	20	42	Establishment	9
18	Expression	20	43	Modifiability	9
19	Proportion	19	44	Coordination	9
20	Innovation	17	45	Cultural Value	9
21	Efficiency	17	46	User Interaction	7
22	Decoration	16	47	Ensure	7
23	Work	16	48	Layering	7
24	Matching	15	49	Connotation	6
25	Clay	15	50	Fusion	6

, “culture” (136 mentions) and “design” (123 mentions) emerged as the most prominent keywords, underscoring a strong practitioner focus on how AIGC can integrate cultural heritage and innovative design. Frequent mentions of “technology” (93) and “Zisha” (92) reveal expectations for deeper integration of modern technological tools with traditional craftsmanship. Within this extended framework, practitioners’ concerns were grouped into instinctive, behavioral, and reflective layers. The instinctive layer highlights immediate sensory appeal—keywords like “appearance” (65 mentions) and “proportion coordination” (19) reflect the importance of visual precision in AIGC-generated designs. The behavioral layer focuses on functionality and real-life applications—evidenced by “practicality” (24 mentions) and “user interaction” (7)—while the reflective layer centers on cultural expression and emotional resonance, with “cultural value” (9 mentions) and “innovation” (17 mentions) indicating a desire to balance heritage preservation and creative innovation. Furthermore, the open-ended questionnaire revealed emerging demands beyond traditional Emotional Design Theory. The additional improvement potential and flexibility layer introduced in the proposed framework addresses these concerns. For instance, “potential for improvement” (41 mentions) and “modifiability” (9 mentions) demonstrate practitioners’ desire for AIGC-generated designs that adapt to evolving needs, enhancing feasibility and market competitiveness. These insights suggest new directions for the evaluation system, acknowledging iterative AIGC features like post-modification and ongoing design improvements.

4.2. Delphi Method: Data Analysis and Results

4.2.1. Expert Analysis

The questionnaire results established a solid foundation for the development of preliminary evaluation indicators. Based on practitioners’ feedback and the word frequency analysis, this study developed an initial evaluation indicator system that includes dimensions such as “appearance appeal”, “functionality and innovation”, “cultural value”, and “improvement potential and flexibility”. This system provides data support and theoretical grounding for the subsequent application of the Delphi method and the Analytic Hierarchy Process. To construct the evaluation indicator system for Yixing Zisha pottery design schemes generated by AIGC technology, the research team conducted two rounds of Delphi consultations with 16 experts selected based on established criteria. All 16 questionnaires were distributed and returned in both rounds, achieving a 100% response rate. The invited experts were all senior practitioners with more than 10 years of professional experience, 75% of whom had backgrounds in Yixing Zisha design or production, ensuring the consultation results were closely aligned with the research objectives. The overall composition of experts was highly representative.

The indicator screening criteria included an average importance score of ≥4.0 and a CV of <0.25. The consultation quality was evaluated across four dimensions: expert enthusiasm, authority, opinion concentration, and coordination. The response rate of the questionnaires reflected expert enthusiasm; authority was determined by the average scores of experts’ judgment bases and familiarity levels. The familiarity levels were scored as follows: very familiar = 1.00, quite familiar = 0.80, moderately familiar = 0.60, slightly familiar = 0.40, not familiar = 0.20. The closer the authority coefficient was to 1, the higher the level of authority. Coordination was assessed using Kendall’s W coefficient, with a significance level of p < 0.05 indicating consensus among expert opinions. The weights of the indicators were subsequently calculated using the Analytic Hierarchy Process.

To ensure the scientific rigor and reliability of data processing, this study utilized Excel 2017 and SPSS 29.0 for data organization and analysis. First, data entry was performed in Excel and verified independently by two researchers to ensure accuracy. Then, statistical analyses were conducted using SPSS, including calculations of mean importance scores, standard deviations, and CVs. Kendall’s W coefficient was also tested for significance to verify the coordination and consistency of expert opinions. These statistical results provided a robust foundation for subsequent weight calculations.

4.2.2. Expert Enthusiasm Coefficient

The expert enthusiasm coefficient typically gauges the reliability of expert consultation. In this study, two rounds of expert consultations were carried out. In the first round, 16 questionnaires were distributed, and all 16 were returned, resulting in a 100.0% return rate, with five experts providing textual suggestions, accounting for 31.25%. In the second round, 16 questionnaires were distributed, and all were returned, maintaining a 100.0% return rate, with three experts offering textual suggestions, making up 18.75%. The effective return rate of documents from both rounds of expert consultations exceeded 90.0%, indicating a high level of expert enthusiasm, as detailed in

Table 4. Analysis of expert enthusiasm coefficient.

Round	Distributed Questionnaires	Collected Questionnaires	Valid Questionnaires	Enthusiasm Coefficient (%)
Round 1	16	16	16	100
Round 2	16	16	16	100

.

4.2.3. Expert Authority Coefficient

In this study, the expert authority coefficient (Cr) was calculated using the following formula:

$$Cr={\displaystyle \frac{Ca+Cs}{2}}$$

where

$Ca$ represents the expert judgment coefficient

$Cs$ represents the expert familiarity coefficient.

The value of $Cr$ was found to be 0.89, which exceeds the recommended threshold of 0.70, indicating a high level of expert authority (see

Table 5. Expert authority coefficient.

Expert	Expert Judgment Coefficient (Ca)	Expert Familiarity Coefficient (Cs)	Expert Authority Coefficient (Cr)
Expert 1	0.90	0.80	0.85
Expert 2	0.80	0.60	0.70
Expert 3	1.00	0.80	0.90
Expert 4	0.80	0.80	0.80
Expert 5	0.90	0.80	0.85
Expert 6	1.00	0.80	0.90
Expert 7	1.00	1.00	1.00
Expert 8	0.90	0.60	0.75
Expert 9	0.90	1.00	0.95
Expert 10	0.90	1.00	0.95
Expert 11	0.80	1.00	0.90
Expert 12	0.90	1.00	0.95
Expert 13	0.90	0.80	0.85
Expert 14	1.00	0.80	0.90
Expert 15	1.00	1.00	1.00
Expert 16	0.90	1.00	0.95
Total	0.91	0.86	0.89

). A Cr value ≥ 0.70 signifies that the expert consultation process is highly reliable, and the expert recommendations can be confidently adopted. To calculate $Ca$ and $Cs$, experts provided self-assessments of their judgment basis and familiarity with the subject matter. This systematic calculation underscores the reliability and rigor of the expert feedback in this study.

4.2.4. Concentration of Expert Opinions

The concentration of expert opinions is reflected in each item’s average importance scores and total mark rates. In the first round of consultation, the average importance scores ranged from 4.125 to 4.938, with an average of 4.484; the total mark rates varied from 31.3% to 93.8%, averaging 58.1%. In the second round, the average importance scores were between 4.125 and 4.875, with an average of 4.455; the total mark rates ranged from 25.0% to 87.5%, averaging 54.2%.

4.2.5. Coordination of Expert Opinions

The coordination of expert opinions throughout the two rounds of consultation demonstrated significant improvement and consistency in the evaluation process. In the first round, the CV values for item importance scores ranged from 0.051 to 0.235, with an average CV of 0.143, reflecting a relatively moderate dispersion of expert judgments. In the second round, the CV values narrowed further, ranging from 0.070 to 0.217, with an average CV of 0.142. The decrease in CV values across the two rounds indicates a convergence of expert opinions as the consultation progressed. Notably, all item CV values in both rounds remained below the threshold of 0.25, which is widely accepted as an indicator of good consistency among expert evaluations. This suggests that the consultation process effectively facilitated a more unified and focused assessment of the evaluation indicators.

Table 6. Inquiry coordination coefficient and significance test results.

Round	Indicator Level	Number of Items	W	Χ2	p
Round 1	Primary Indicators	4	0.111	5.349	0.148
	Secondary Indicators	16	0.134	32.266	0.006
	Overall	20	0.126	38.398	0.005
Round 2	Primary Indicators	4	0.112	5.400	0.145
	Secondary Indicators	17	0.202	51.682	<0.001
	Overall	21	0.206	65.911	<0.001

illustrates the coordination coefficient (Kendall’s W) results, which serve as a measure of the agreement level among the panel of experts. In the first round, Kendall’s W was calculated as W = 0.126 (p = 0.005, significant at p < 0.01), indicating moderate agreement at the initial stage. By the second round, Kendall’s W increased to W = 0.206 (p < 0.001), signifying a higher degree of consistency among the experts’ responses. The substantial improvement in Kendall’s W between rounds further underscores the effectiveness of the Delphi process in harmonizing expert opinions.

4.2.6. First Delphi Survey

In the first round of expert consultation, experts were asked to evaluate the importance of each item on the questionnaire based on their practical experience, theoretical knowledge, references to domestic and international materials, and personal subjective feelings. The analysis of the results is as follows: In this round, the average scores, standard deviations, coefficients of variation, and total mark rates for primary and secondary indicators are shown in

Table 7. First-round Delphi consultation result.

Measure	Mean (M)	Standard Deviation (SD)	Coefficient of Variation (CV)	Full Score Rate (%)	Non-Conformities	Result
Instinctual Level	4.813	0.403	0.084	81.30%	0	Retain
Appearance Attractiveness	4.875	0.342	0.07	87.50%	0	Retain
Color Coordination	4.125	0.885	0.215	43.80%	0	Retain
Material Texture	4.438	0.727	0.164	56.30%	0	Retain
Texture and Decoration Accuracy	4.563	0.512	0.112	56.30%	0	Retain
Aesthetic Form	4.938	0.25	0.051	93.80%	0	Retain
Proportional Harmony	4.688	0.602	0.128	75.00%	0	Retain
Behavioral Level	4.313	0.946	0.219	56.30%	0	Retain
Practicality	4.25	1	0.235	56.30%	0	Retain
User Interaction Friendliness	4.188	0.655	0.156	31.30%	0	Retain
Functionality and Innovation	4.5	0.516	0.115	50.00%	0	Retain
Reflective Level	4.563	0.629	0.138	62.50%	0	Retain
Creativity and Originality	4.375	0.806	0.184	56.30%	0	Retain
Cultural Heritage Expression	4.5	0.632	0.141	56.30%	0	Retain
Cultural Identity	4.438	0.727	0.164	56.30%	0	Retain
Cultural Value	4.438	0.629	0.142	50.00%	0	Retain
Improvement Potential and Flexibility	4.438	0.512	0.115	43.80%	0	Retain
Design Modifiability	4.563	0.512	0.112	56.30%	0	Retain
Feasibility	4.25	0.775	0.182	43.80%	0	Retain
Enhanced Value Post-Modification	4.438	0.629	0.142	50.00%	0	Retain

. The items under consideration were 4 primary and 16 secondary items, totaling 20. The average importance scores for these items ranged from 4.125 to 4.938, with coefficients of variation between 0.051 and 0.235 and total mark rates from 31.3% to 93.8%. The selection criteria were as follows: average score ≥ 3.50, coefficient of variation ≤ 0.25, and total mark rate > 0.0%. Furthermore, additional content suggested by experts, including “creative shape”, was added. The new questionnaire, comprising 4 primary items and 17 secondary items, was prepared for investigation in the second round of consultation.

4.2.7. Results of the Second Round of Expert Consultation

As presented in

Table 8. Second-round Delphi consultation result.

Measure	Mean (M)	Standard Deviation (SD)	Coefficient of Variation (CV)	Full Score Rate (%)	Non-Conformities	Result
Instinctual Level	4.75	0.447	0.094	75.00%	0	Retain
Appearance Attractiveness	4.875	0.342	0.07	87.50%	0	Retain
Color Coordination	4.188	0.544	0.13	25.00%	0	Retain
Material Texture	4.188	0.75	0.179	37.50%	0	Retain
Texture and Decoration Accuracy	4.438	0.629	0.142	50.00%	0	Retain
Aesthetic Form	4.625	0.5	0.108	62.50%	0	Retain
Proportional Harmony	4.625	0.5	0.108	62.50%	0	Retain
Creative Form	4.625	0.619	0.134	68.80%	0	Retain
Behavioral Level	4.563	0.727	0.159	68.80%	0	Retain
Practicality	4.125	0.806	0.195	31.30%	0	Retain
User Interaction Friendliness	4.188	0.911	0.217	43.80%	0	Retain
Functionality and Innovation	4.25	0.775	0.182	43.80%	0	Retain
Reflective Level	4.563	0.512	0.112	56.30%	0	Retain
Creativity and Originality	4.563	0.629	0.138	62.50%	0	Retain
Cultural Heritage Expression	4.563	0.512	0.112	56.30%	0	Retain
Cultural Identity	4.188	0.655	0.156	31.30%	0	Retain
Cultural Value	4.313	0.704	0.163	43.80%	0	Retain
Improvement Potential and Flexibility	4.5	0.816	0.181	68.80%	0	Retain
Design Modifiability	4.625	0.5	0.108	62.50%	0	Retain
Feasibility	4.188	0.75	0.179	37.50%	0	Retain
Enhanced Value Post-Modification	4.625	0.5	0.108	62.50%	0	Retain

, the second round of expert consultation evaluated a total of 21 items, comprising 4 primary indicators and 17 secondary indicators. The average importance scores for these items ranged from 4.125 to 4.875, with coefficients of variation between 0.070 and 0.217, and total mark rates spanning from 25.0% to 87.5%. The selection criteria for indicators were set as follows: average scores ≥ 3.50, CV ≤ 0.25, and total mark rates > 0.0%. In the second round of expert consultation, the overall results indicated that, following further expert feedback and optimization, the stability and consistency of the evaluation system were significantly enhanced. Although some minor discrepancies in ratings were observed for specific indicators such as material texture and color coordination, overall, the experts’ evaluations became more consistent and met the selection criteria. The mean values for both primary and secondary indicators remained high, demonstrating the experts’ consistent recognition of dimensions such as visual appeal, functional innovation, and cultural expression.

These changes demonstrate that through two rounds of expert feedback, the evaluation system has been further optimized, providing a solid foundation for the subsequent weight calculations and decision analysis using the AHP. At the same time, experts expressed positive feedback on the potential of AIGC technology in Yixing Zisha design, validating its capacity to enhance both design innovation and cultural heritage preservation.

4.3. AHP Analysis and Results

4.3.1. Construction of the Hierarchical Structure Model

Based on two rounds of Delphi expert consultations—and according to the experts’ evaluations in the AIGC-generated Yixing Zisha design evaluation system—a hierarchical analysis model was constructed using Yaahp software Version 11.0. As shown in Figure 6, this study’s decision objective is to evaluate AIGC-generated Yixing Zisha teapot designs, with 4 primary indicators at the element level and 17 secondary indicators at the alternative solution level. This process reinforced the scientific validity of the 4 primary and 17 secondary indicators determined by Delphi, clarifying their relative importance and offering a solid basis for a rigorous and systematic evaluation framework.

4.3.2. Construction of Judgment Matrices

Based on the hierarchical structure model of the evaluation indicator system for Yixing Zisha teapot design schemes generated with AIGC technology, the Saaty scale method was applied to conduct pairwise comparisons among the indicators at each level, thereby constructing judgment matrices. The mean differences (ΔZ) of the importance ratings assigned by experts were used to determine the corresponding Saaty scale values. The Saaty scale and its explanations are shown in

Table 9. Saaty scale table.

ΔZ Value	Degree	Saaty Scale
ΔZ = 0	Equal importance	1
0 < ΔZ ≤ 0.25	Slightly more important	2
0.25 < ΔZ ≤ 0.50	Moderately more important	3
0.50 < ΔZ ≤ 0.75	More important	4
0.75 < ΔZ ≤ 1.0	Much more important	5
1.0 < ΔZ ≤ 1.25	Significantly more important	6
1.25 < ΔZ ≤ 1.5	Very significantly more important	7

Note: ΔZ represents the mean difference in importance scores between two indicators (α, β). For example, when 0.25 < ΔZ ≤ 0.50, it indicates that α is moderately more important than β, corresponding to a Saaty scale value of 3. Conversely, if −0.50 < ΔZ ≤ −0.25, it indicates that β is moderately more important than α, corresponding to the reciprocal scale value of 1/3, and so on.

.

The judgment matrix for the primary indicators, constructed based on expert opinions, is presented in

Table 10. Primary indicator judgment matrix.

Primary Indicator	Instinctive Layer	Behavioral Layer	Reflective Layer	Improvement Potential and Flexibility
Instinctive Layer	1	2	2	2
Behavioral Layer	1/2	1	1	2
Reflective Layer	1/2	1	1	2
Improvement Potential and Flexibility	1/2	1/2	1/2	1

. This matrix illustrates the pairwise comparisons and relative importance among the four primary indicators: instinctive level, behavioral level, reflective level, and improvement potential and flexibility. This matrix quantitatively represents the relative importance of the primary indicators, serving as the foundation for subsequent weight calculations and consistency checks. Based on the above method, the judgment matrices for the secondary indicators were subsequently derived.

4.3.3. Indicator Weights and Consistency Validation

Building upon the constructed judgment matrices, this study utilized Yaahp software to perform quantitative analysis of indicator weights and conduct rigorous consistency checks. The results demonstrated logical and scientifically valid weight distributions for both primary and secondary indicators, with all CR values below the 0.1 threshold. This confirms the reliability and coherence of the evaluation framework, providing a strong basis for subsequent research and practical applications in assessing AIGC-generated designs.

The weight distribution for the four primary indicators is summarized in

Table 11. Primary indicator weights and consistency validation.

Primary Indicator	Weight	CR	λmax
I-1 Instinctive Layer	0.3952	0.0227	4.0606
I-2 Behavioral Layer	0.2322
I-3 Reflective Layer	0.2322
I-4 Potential for Improvement and Flexibility	0.1404

. The instinctive layer holds the highest weight at 0.3952, highlighting the critical role of visual and sensory appeal in the evaluation of AIGC-generated designs, as it directly impacts user engagement and aesthetic satisfaction. The behavioral layer and reflective layer share equal weights at 0.2322, indicating the balanced importance of functional usability and cultural resonance. These dimensions ensure that the designs are not only practical but also align with the cultural values inherent in Yixing Zisha pottery. The improvement potential and flexibility dimension carries the lowest weight at 0.1404, emphasizing its supplementary role in capturing the adaptability and potential for enhancement of design schemes.

Consistency validation of the primary indicator weights yielded a CR of 0.0227, which is well below the 0.1 threshold defined by the AHP methodology. This result reaffirms the logical consistency and scientific rigor of the evaluation system, ensuring its robustness for application in both academic research and practical evaluations.

Based on the results from the AHP model and the judgment matrices, the weights and consistency check for the secondary indicators were calculated and analyzed. The specific results are shown in

Table 12. Weights and consistency test results for secondary indicators in the evaluation system.

Primary Indicator	Secondary Indicator	Weight	Combined Weight	CR	λmax
Instinctive Layer	Appearance Attractiveness	0.2888	0.1142	0.0052	7.0422
	Color Coordination	0.0570	0.0225
	Material Texture	0.0570	0.0225
	Texture and Decoration Accuracy	0.0952	0.0376
	Aesthetic Form	0.1674	0.0661
	Proportional Harmony	0.1674	0.0661
	Creative Form	0.1674	0.0661
Behavioral Layer	Practicality	0.1958	0.0455	0.0516	3.0536
	User Interaction Friendliness	0.3108	0.0722
	Functionality and Innovation	0.4934	0.1146
Reflective Layer	Creativity and Originality	0.3509	0.0815	0.0039	4.0104
	Cultural Heritage Expression	0.3509	0.0815
	Cultural Identity	0.1091	0.0253
	Cultural Value	0.1891	0.0439
Improvement Potential and Flexibility	Design Modifiability	0.4286	0.0602	0.0000	3.0000
	Feasibility	0.1429	0.0201
	Enhanced Value Post-Modification	0.4286	0.0602

. The weights for the secondary indicators across the four primary levels (instinctive layer, behavioral layer, reflective layer, and improvement potential and flexibility) vary, reflecting the importance assigned to each attribute in the evaluation system. For the instinctive layer, the indicator “appearance attractiveness” (0.2888) received the highest weight, emphasizing the critical role of visual appeal in evaluating AIGC-generated designs. Other secondary indicators such as “aesthetic form”, “proportional harmony”, and “creative form” all share the exact weight of 0.1674, indicating that these aspects of design carry similar importance in the evaluation. Conversely, “color coordination” and “material texture” were assigned the lowest weight (0.0570), highlighting their relatively lesser impact in this dimension. The CR of 0.0052 and λmax of 7.0422 demonstrate a high level of consistency in the expert judgments for this layer. In the behavioral layer, the “functionality and innovation” indicator had the highest weight of 0.4934, underscoring the importance of practical functionality and innovative aspects in AIGC-generated Yixing Zisha designs. “User interaction friendliness” followed with a weight of 0.3108, and “practicality” had the lowest weight of 0.1958, reflecting the varying significance attributed to usability and design interaction. The CR value of 0.0516 and λmax of 3.0536 indicate satisfactory consistency in the expert evaluations for this layer. The reflective layer highlighted the cultural and emotional aspects of the design. Both “creativity and originality” and “cultural heritage expression” received the highest weight of 0.3509, emphasizing the centrality of creative innovation and cultural significance in the design evaluation. “Cultural value” and “cultural identity” received comparatively lower weights of 0.1891 and 0.1091, respectively, reflecting the relative importance of cultural aspects in the broader evaluation framework. The CR value of 0.0039 and λmax of 4.0104 suggest a high degree of consensus among the experts in evaluating these reflective aspects. For the improvement potential and flexibility layer, “design modifiability” and “enhanced value post-modification” both received the highest weight of 0.4286, indicating the experts’ strong emphasis on the adaptability and optimization potential of the AIGC-generated designs. “Feasibility” was assigned the lowest weight of 0.1429, but it still played a crucial role in the overall evaluation. The CR value of 0.0000 and λmax of 3.0000 indicate perfect consistency in the judgment matrix for this layer.

The weights of the secondary indicators relative to the overall indicators range from 0.0201 to 0.1146, with CR for all secondary indicators being less than 0.1. The results show that the allocation of weights for the secondary indicators is reasonable, and the consistency test results meet the required conditions, making them suitable for decision-making.

Figure 7 presents the evaluation indicator weights for the Yixing Zisha teapot design schemes generated by AIGC technology. The comprehensive weight distribution across the evaluation indicators reflects a balanced approach that emphasizes the aesthetic, functional, and cultural attributes of AIGC-generated designs for Yixing Zisha teapots. These results underscore the importance of maintaining traditional visual appeal while integrating innovation and cultural depth. High weights for aesthetic and cultural indicators reveal a strong alignment with the unique values inherent in Yixing Zisha pottery. Meanwhile, the focus on flexibility and adaptability in the improvement potential and flexibility layer suggests that experts value designs capable of evolving with user needs and technological advancements.

5. Discussion

5.1. Innovativeness of the Evaluation Indicator System

This study’s evaluation indicator system brings AIGC technology into dialogue with the traditional Yixing Zisha pottery craft, building on Norman’s Emotional Design Theory while addressing the unique characteristics of AIGC-driven processes. Particularly, the inclusion of “improvement potential and flexibility” acknowledges iterative design workflows that emerged from expert feedback and Delphi–AHP validations. In this sense, the system broadens the application of Emotional Design Theory in a heritage context, highlighting new demands of AIGC-driven design.

Integrating Emotional Design Theory with intangible cultural heritage marks a key theoretical contribution. Norman (2004) emphasized the importance of the instinctive, behavioral, and reflective levels in design, focusing on emotional responses. This study extends this theoretical foundation by introducing a multidimensional evaluation framework that incorporates an additional layer—improvement potential and flexibility—capturing the iterative and adaptive nature of AIGC technology. Adapting Norman’s model, this study evaluates Yixing Zisha designs generated through AIGC, emphasizing the cultural resonance and emotional bonds inherent in heritage crafts [39]. This ensures the evaluation transcends surface aesthetics and functionality, delving into deeper cultural connotations. Norman’s theory thus offers a framework for understanding user perception and emotional responses, which is especially apt for traditional craft designs. It captures the users’ need for cultural identity and artistic value—now considered essential in AIGC-generated outputs.

Moreover, this study also demonstrates how AIGC technology can enrich Yixing Zisha pottery design processes. AIGC is not merely a design tool but serves as a bridge between traditional craftsmanship and modern methods. Compared to Wang and Zhan (2024), which mainly focuses on design efficiency and innovation from AIGC, this study includes “cultural heritage expression” among core indicators, ensuring that AIGC not only enhances efficiency but also retains cultural and symbolic meanings in Yixing Zisha pottery [34]. This approach facilitates a harmonious coexistence of tradition and innovation.

At the instinctive level, indicators such as “appearance attractiveness” (weight 0.2888) and “aesthetic form” (0.1674) highlight the importance of visual harmony in capturing users’ initial impressions. At the reflective level, indicators such as “cultural heritage expression” and “creativity and originality” (both 0.3509) underline the importance of long-term emotional attachment. These levels, as outlined in the theoretical framework, ensure that evaluation captures both immediate aesthetic impressions and deeper emotional resonance. These findings resonate with He (2022), who explored how artificial intelligence enhances the cultural authenticity of ceramic designs [2]. While He’s work emphasizes creative efficiency, this study extends the scope by demonstrating that reflective-level indicators like “cultural identity” (0.1091) and “cultural value” (0.1891) are equally crucial for preserving emotional resonance in AIGC-generated design. In addition to enriching Emotional Design Theory, this study introduces the novel dimension “improvement potential and flexibility”. As practitioners emphasized, iterative workflows are integral to refining texture details, adjusting proportions, and integrating eco-friendly materials. Indicators such as “design modifiability” and “enhanced value post-modification”, which received high weights (e.g., 0.4286), ensure that the evaluation system accommodates the adaptive and generative capabilities of AIGC. This dimension, reflected in the theoretical framework, not only addresses iterative design needs but also ensures the model’s relevance to evolving market dynamics and heritage preservation goals. This approach echoes the observations of Xin Tian et al. (2022), who identified adaptability and functionality as critical in AIGC-driven industrial design processes [62]. However, while Tian’s framework focused on functionality in industrial products, this study integrates cultural dimensions, ensuring its applicability to heritage-rich contexts like Yixing Zisha pottery.

By grounding the evaluation indicator system within the theoretical framework, this study reconciles technological adaptability with heritage authenticity. The multidimensional framework—combining instinctive, behavioral, reflective, and AIGC-oriented features—holistically assesses the cultural and practical value of AIGC-generated designs, providing a structured tool for balancing cultural preservation with innovation.

5.2. Cultural Heritage Preservation and Sustainable Development

The application of AIGC technology in Yixing Zisha pottery design illustrates how modern tools can empower traditional crafts while preserving their cultural essence. Prasad et al. (2024) emphasized how digital technologies enhance the dissemination of traditional culture and stimulate creativity [63]. Similarly, Mendoza et al. (2023) argued that AIGC-based innovations enable new pathways for engaging global audiences while preserving the symbolic meaning of heritage crafts [21]. This study builds on their insights, showing how reflective-layer indicators—e.g., “cultural heritage expression” (weight 0.3509)—maintain symbolic and emotional worth in Yixing Zisha pottery across modern markets.

Additionally, AIGC fosters sustainable heritage preservation via iterative design processes that minimize material waste. As the Delphi panel suggested, artisans can digitally prototype shapes and patterns prior to physical crafting, thereby reducing resource use. High weights for sub-indicators such as “enhanced value post-modification” (0.4286) underscore the significance of sustainability, aligning with global conservation efforts that call for balancing innovation with preservation. While “feasibility” scored relatively lower (0.1429), experts expressed confidence in AIGC’s technical viability—shifting focus to refining cultural and ecological value.

This study further highlights how AIGC-based designs respond to evolving market demands while retaining cultural essence. Transforming traditional decorative motifs into innovative concepts, AIGC paves new avenues for cultural innovation and broader audience reach. Future research can extend this framework to other traditional crafts, thereby advancing the dual objectives of cultural innovation and preservation.

5.3. Deepening the Quantitative Analysis of Findings

The Delphi–AHP results offer nuanced insights into the priorities across the three levels of Emotional Design Theory, shedding light on the preferences and evolving considerations of practitioners. At the instinctive level, indicators such as “appearance attractiveness” (0.2888) underscore the importance of visual appeal in capturing consumer interest, reaffirming the longstanding emphasis on the artistic identity of Yixing Zisha pottery. In contrast, indicators such as “material texture” (0.0570) received lower weights, reflecting the relatively limited emphasis on tactile qualities compared to visual elements in initial user engagement.

In the behavioral realm, “functionality and innovation” emerged as the most significant indicator (weight 0.4934), demonstrating the importance of aligning traditional craftsmanship with modern usability and market relevance. This finding suggests that the economic viability of Yixing Zisha pottery in contemporary markets depends on its ability to balance traditional craftsmanship with practical innovation.

At the reflective level, “cultural heritage expression” and “creativity and originality” (both 0.3509) jointly emphasize the dual importance of authenticity and novelty in fostering prolonged user engagement. Sub-indicators such as “cultural value” (0.1891) and “cultural identity” (0.1091) further illustrate the pivotal role intangible cultural elements play in shaping user perceptions and acceptance of AIGC-generated designs.

By situating these findings within the broader cultural, historical, and technological contexts, this study presents a holistic perspective on how AIGC can reimagine Yixing Zisha pottery for modern audiences. The convergence of empirical evidence with theoretical insights not only enriches our understanding of intangible cultural heritage preservation but also highlights practical strategies for integrating advanced technologies into traditional craftsmanship.

6. Conclusions

Yixing Zisha pottery, as a significant intangible cultural heritage representing traditional Chinese craftsmanship, confronts the dual imperative of preserving its cultural essence while innovating to meet evolving consumer demands. The advent of AIGC technology provides a potent instrument for this transformation, facilitating the creation of novel design schemes while respecting deeply rooted cultural values. This study develops a comprehensive and sustainable evaluation framework for AIGC-generated Yixing Zisha pottery design schemes, explicitly rooted in data-driven insights and validated through Delphi and AHP methods. By integrating Emotional Design Theory with iterative dimensions, the framework bridges traditional craftsmanship with modern digital tools, offering a replicable model adaptable to other intangible cultural heritage domains undergoing digital transformation.

The study’s findings emphasize the critical roles of visual appeal and cultural resonance in evaluating AIGC-generated designs. Key indicators, such as “appearance attractiveness” (0.2888) and “cultural heritage expression” (0.3509), emerged as pivotal elements based on practitioners’ feedback and expert validation. This demonstrates that the framework is firmly grounded in empirical data, aligning practitioners’ priorities with theoretical constructs. High weighting of instinctive indicators—such as appearance attractiveness and proportional harmony—reveals the enduring importance of visual and sensory elements in Yixing Zisha pottery, confirming that designs must retain the craft’s traditional aesthetic values. Meanwhile, reflective indicators—like cultural heritage expression and creativity—underscore cultural authenticity and emotional attachment, demonstrating that these designs must not only be visually appealing but also meaningfully aligned with their cultural origins. A significant theoretical innovation within this framework is the “improvement potential and flexibility” dimension, directly informed by data analysis, acknowledging that AIGC technology enables iterative, adaptive workflows far beyond static craft approaches. Indicators such as “design modifiability” and “enhanced value post-modification”—both weighted highly—reflect practitioners’ desire for continuous refinement of shapes, textures, and environmental footprints. This dimension addresses forward-looking aspects of AIGC technology, broadening the conventional application of Emotional Design Theory, which typically highlights instinctive, behavioral, and reflective layers.

From a theoretical perspective, this research demonstrates that Emotional Design Theory can be productively applied to intangible cultural heritage contexts where symbolic meaning, historical narratives, and emotional identification drive consumer engagement. By integrating functional performance with cultural authenticity and introducing adaptive design features, this study broadens the theory’s scope to accommodate the iterative and dynamic nature of AIGC processes. This framework emphasizes the dual imperatives of creative efficiency and cultural preservation, addressing a critical gap in the existing literature. On a practical level, the research contributes valuable insights for artisans, designers, and policymakers. The evaluation framework offers clear, actionable guidelines for embedding AIGC in traditional crafts, assisting artisans in refining creative processes to ensure cultural integrity while introducing modern functionalities. For policymakers, the framework provides a foundation for setting standards, funding priorities, and training programs aimed at sustainable innovation in the creative industries. The emphasis on iterative design and ecological sustainability ensures that the framework aligns with global sustainability goals, offering tools for material efficiency and reduced waste in traditional crafts. Moreover, the research underscores the sustainability potential of AIGC technology in traditional crafts. By facilitating iterative design processes, AIGC can reduce material waste, optimize resource usage, and enhance environmental responsibility—objectives increasingly vital in global sustainability efforts. This study highlights the potential for cross-cultural collaborations, demonstrating how AIGC-enabled designs can promote heritage crafts like Yixing Zisha pottery in diverse consumer landscapes while maintaining their symbolic and ecological values.

In conclusion, the evaluation framework articulated in this study addresses the twin goals of protecting cultural heritage and promoting technological innovation. It offers a scientifically validated model for bridging tradition with modernity, ensuring that intangible cultural heritage is preserved and enhanced within rapidly evolving digital ecosystems. Future studies could extend this framework by examining user-centric testing, generational preferences, or cross-cultural comparisons, further refining how intangible heritage can adapt to fast-changing technological landscapes while retaining its cultural soul.