Research on AIGC-Integrated Design Education for Sustainable Teaching: An Empirical Analysis Based on the TAM and TPACK Models

Abstract

With the rapid proliferation of Generative Artificial Intelligence (AIGC) technologies in higher education, identifying effective integration pathways into design curricula has become a pressing issue in the field of educational technology. This study employs a mixed-methods approach grounded in the Technology Acceptance Model (TAM) and the Technological Pedagogical Content Knowledge (TPACK) framework, incorporating course analysis, questionnaire surveys, structural path modeling, and interview analysis. Focusing on both instructors and students, this research systematically investigates the acceptance, integration mechanisms, and sustainable development potential of AIGC in university-level design education. The findings indicate that students generally acknowledge the value of AIGC in enhancing creativity and improving efficiency, although gaps persist in their understanding of ethical considerations and original expression. On the teachers’ side, technological knowledge exerts a significant positive influence on the integration of content knowledge, while the impact of pedagogical knowledge remains underutilized. Interview data further reveal a structural tension within current teaching practices, characterized by the rapid adoption of technological tools contrasted with the slower evolution of pedagogical systems. Based on these insights, this study offers five strategic recommendations for sustainable teaching, including the development of teacher training systems, curriculum module design, student media literacy enhancement, and pedagogical reconstruction aligned with the Sustainable Development Goals (SDGs). These findings provide both theoretical and practical contributions to the effective and sustainable integration of AIGC into higher design education.

1. Introduction

In recent years, Generative Artificial Intelligence (AIGC) technologies have undergone rapid development and found widespread applications in content creation, artistic design, and educational training, drawing sustained attention from both academia and the education sector [1]. Particularly within educational settings, the growing popularity of AIGC tools such as ChatGPT, Midjourney, and Stable Diffusion is profoundly reshaping traditional teaching models and pathways of knowledge acquisition [2]. Against this backdrop, design education—which emphasizes creative expression and visual thinking—has emerged as a key frontier for the application and experimentation of AIGC technologies [3]. In this study, the term AIGC (AI-Generated Content) refers not only to the output of generative AI technologies (such as images or texts), but also to the technological process and tools (e.g., ChatGPT, Midjourney, Stable Diffusion) that enable automated creation. While “GenAI” is sometimes used to refer to the technology itself, AIGC has become a widely accepted term in Chinese academic discourse, especially in design and content-innovation contexts. Thus, for consistency with local terminology and relevance to educational applications, we use “AIGC” throughout this paper.

With their capabilities in illustration creation, image generation, and style transfer, AIGC tools have the potential to enhance students’ creative efficiency and expressive abilities. However, they also introduce new challenges to course structures, instructional processes, and assessment systems [4]. On the one hand, AIGC brings a new paradigm of “automated generation + creative collaboration” into design education; on the other hand, balancing technological assistance with students’ creative autonomy and reconstructing learner-centered pedagogical approaches have become central concerns in ongoing curricular reforms within higher education institutions.

In Chinese universities, the educational application of AIGC technologies has already established a preliminary foundation. Some design schools have integrated AIGC into courses such as illustration, visual communication, and product design. Although existing studies have begun to explore the educational potential of AIGC, systematic empirical research on its integration into design curricula remains limited [5,6]. In particular, from the perspective of “sustainable teaching transformation,” the deeper impacts of AIGC integration on course structure optimization and pedagogical innovation mechanisms have yet to be fully revealed [7].

Therefore, this study takes design courses in Chinese universities as its research context and constructs an analytical framework based on the Technology Acceptance Model (TAM) and the Technological Pedagogical Content Knowledge (TPACK) model. Using course content analysis, questionnaire surveys, and interview investigations, this study attempts to gain an in-depth understanding of the current status of AIGC integration in higher education design teaching, including the levels of acceptance among teachers and students, as well as the major challenges faced in implementation. Based on the findings, it proposes strategic suggestions aimed at facilitating a more sustainable transformation in teaching.

This study is structured around the following three core research questions:

RQ1: What is the current status of AIGC tool application in design courses at Chinese universities? In which types of courses and teaching stages are these tools primarily utilized?

RQ2: What are the similarities and differences between teachers’ and students’ perceptions, attitudes, and acceptance of AIGC integration in teaching? What are the main influencing factors?

RQ3: What are the major issues and challenges currently faced in AIGC-assisted teaching practices? How can existing teaching models be optimized to improve integration effectiveness?

Based on the above questions, this study establishes the following specific research objectives:

• To map the current integration patterns of AIGC technologies within representative design courses in Chinese universities, and to identify the main instructional structures, tool application stages, and curriculum types involved.
• To compare and analyze the acceptance factors of AIGC-integrated teaching from the perspectives of both teachers and students by applying the TAM and TPACK models, focusing on the influence of technological, pedagogical, and emotional dimensions on behavioral intention.
• To reveal structural gaps and developmental bottlenecks in current AIGC teaching practices and to propose strategic pathways for sustainable curriculum transformation aligned with pedagogical goals and the Sustainable Development Goals (SDGs).

To address these research questions, this study adopts a mixed-methods empirical approach that combines curriculum content analysis, structured questionnaires, and semi-structured interviews. From the perspectives of both teachers and students, this study explores the key factors influencing AIGC integration in design education, including the cognitive, emotional, and ethical dimensions of technology acceptance. Theoretically, it is grounded in the TAM and the TPACK model, aiming to provide an analytical framework that supports sustainable curriculum innovation and pedagogical reform in higher education.

2. Literature Review

2.1. The Application and Development of AIGC in Design Education

Since the introduction of the Transformer model in 2017, these AI technologies have advanced rapidly in generating complex, diverse, and efficient content. Representative tools such as ChatGPT and Midjourney have gradually penetrated into both artistic creation and educational practice. In the context of design education, AIGC is increasingly being adopted across multiple stages, including idea generation, style transfer, and prototype development, thereby enabling more intelligent and personalized teaching processes.

Previous studies have shown that the introduction of AIGC contributes to enhancing students’ learning efficiency and classroom engagement. For instance, Xiong (2022) employed convolutional neural networks (CNNs) to evaluate the teaching effectiveness of visual communication courses, which significantly improved classroom interaction levels [8]. Zheng (2024) developed a system based on adaptive learning algorithms, which optimized learning pathways through behavioral data analysis and enhanced the potential for personalized instruction [9]. In addition, Basarir (2022) constructed an AI-centered instructional model for design education, which effectively improved students’ understanding and control of the design process [10].

The openness and collaborative capabilities of technological platforms have also attracted considerable attention. Maselli et al. (2022) proposed the concept of “diffuse design,” emphasizing how AIGC tools can facilitate shared and democratized design processes through collaboration between professional designers and non-expert users [11]. Schleiss et al. (2023) integrated design thinking principles into AI curriculum frameworks, developing an interdisciplinary pedagogical structure that offers theoretical guidance for the systematic incorporation of AIGC into educational settings [12]. Weng et al. (2024) conducted a systematic review of educational practices that integrate AI and computational thinking (CT), finding strong synergies between AI tools and CT methodologies in STEAM education [13]. These studies underscore the convergence of technology and educational innovation, laying a foundation for the development of sustainable curriculum optimization strategies based on generative AI. At the same time, they also highlight the potential challenges inherent in the integration of emerging technologies and curriculum design.

Despite the vast potential demonstrated by AIGC, scholars have also pointed out several challenges related to ethical standards, technological controllability, and curriculum adaptability. For example, Fareed et al. (2024), using heritage architecture education as a case study, noted that, while AIGC enhances immersive learning and visualization capabilities, it also raises concerns about authenticity and algorithmic bias [14]. Thomson et al. (2024) further emphasized the importance of ethical governance in the application of AIGC within educational settings [15]. Tang et al. (2022), building on the TPACK model, proposed new strategies for design education [16]. Their research highlights that AI-enhanced instruction and personalized project-based learning significantly improve students’ learning experiences and design practice capabilities.

In summary, while these studies have revealed multiple application paths of AIGC in design education—spanning tool empowerment, platform collaboration, and interdisciplinary restructuring—they largely remain focused on isolated teaching cases or technology demonstrations. There remains a lack of systematic understanding regarding how AIGC is comprehensively embedded into instructional structures, particularly in higher education contexts. Furthermore, most research does not explicitly align AIGC teaching practices with theoretical models such as TAM and TPACK, nor do they address how these integrations can support sustainable and scalable pedagogical transformation. This section contextualizes the technological and pedagogical foundations of AIGC integration in Chinese design education, thereby directly addressing RQ1 of this study.

2.2. Design Thinking and Educational Innovation Practices

Design thinking, as a human-centered innovation methodology, has shown increasing relevance in educational transformation. Its five-stage process—empathize, define, ideate, prototype, and test—offers structured guidance for addressing pedagogical challenges brought by the integration of emerging technologies such as AIGC. Melles et al. (2012), through the utilization of teaching practices in Australia and Hong Kong, found that open-ended and experiential learning approaches—such as project-based learning and workshop teaching—contributed to enhancing students’ problem-solving abilities [17]. Chao-Ming Yang and Tzu-Fan Hsu (2020) introduced design thinking into a packaging design course, which significantly improved students’ creative self-efficacy and flow experience, thereby increasing their design motivation and learning outcomes [18].

Recent discussions have moved beyond student-centered benefits to highlight how design thinking principles can help educators reframe their instructional strategies in response to AI tools. The transformation of the teacher’s role has also become a prominent topic in current research. The International Society for Technology in Education (ISTE, 2016) proposed that future educators should assume the role of “designers,” not only delivering knowledge, but also systematically constructing curricula to adapt to the evolving landscape of educational technologies [19]. Zhao et al. (2024), in the context of AI empowerment, argued that the integration of design thinking and intelligent technologies will further promote curriculum restructuring and pedagogical transformation [20].

In addition to its value in fostering creativity and learner-centered instruction, design thinking has been increasingly discussed as a conceptual link between emerging technologies and pedagogical adaptation. It emphasizes not only ideation and empathy in the learning process, but also iterative experimentation—an aspect particularly relevant when integrating complex AI tools such as AIGC into curriculum design.

However, most current studies focus on course-level teaching practices or student learning motivation, rather than providing a clear competency framework for teachers tasked with implementing AIGC-enabled instruction. There remains a notable lack of systematic research on how educators can effectively align technological innovation with pedagogical strategies and content knowledge—issues central to the TPACK model. This limitation is particularly evident in contexts where design instructors must quickly adapt to AI-driven tools without sufficient training or structural pedagogical support. Therefore, this section lays the conceptual foundation for investigating the challenges and potentials of AIGC curriculum integration from the educator’s perspective.

2.3. The TPACK Model and Its Application in Design Education

The TPACK model, developed from Shulman’s Pedagogical Content Knowledge (PCK) framework, emphasizes that effective teaching requires an integrated understanding of Technological Knowledge (TK), Pedagogical Knowledge (PK), and Content Knowledge (CK). In the context of AI integration in education, the TPACK model has been widely used to assess teachers’ abilities to effectively incorporate emerging technologies [21].

Research has demonstrated that the TPACK framework is well-suited for application in design education. Tang et al. (2022) [16] introduced the TPACK model into AI-assisted design instruction and found that it significantly improved students’ learning experiences and creative capabilities. From the perspective of teacher development, Karataş and Ataç (2024) reported that, although preservice teachers generally possess basic technological knowledge, their competence in AI integration and awareness of ethical implications still require enhancement [22].

In recent years, the TPACK model has undergone several notable extensions. Lan et al. (2025) proposed the GenAI-TPACK model, which incorporates Technological Ethical Assessment Knowledge (TEAK) to address the need for ethical risk management in AIGC-based teaching [23]. Celik (2023) [6] introduced the Intelligent-TPACK model, which emphasizes dimensions such as fairness, transparency, and accountability in technology integration, thus providing a more comprehensive framework for evaluating teachers’ AI teaching competencies. Farangi, Nejadghanbar, and Hu (2024) [24] provide additional insight into the ethical challenges of integrating generative AI into curriculum design [24]. Their study emphasizes the need for anticipatory ethical governance at the pedagogical planning stage and highlights the role of educator training in mitigating unintended consequences. This aligns with the ethical dimension of the TEAK framework and supports the inclusion of “ethical perception” as a contextual factor in this study’s extended TAM.

While these contributions enrich the theoretical dimensions of TPACK, most existing studies focus on its conceptual development or general applicability in technology-enhanced learning. There is still a lack of empirical research that integrates TPACK with AIGC-specific teaching practices in higher design education, especially in large-scale curriculum reform contexts. Moreover, few studies have quantitatively examined the interrelationships among TK, PK, and CK in real instructional settings, leaving the dynamic interaction among these components underexplored.

To bridge this gap, this study explores the impact of TK and PK on CK by using an empirical analysis based on teacher questionnaires and regression modeling. Such an approach allows for a deeper understanding of how design instructors perceive and operationalize their professional knowledge in response to AIGC integration.

To summarize, the TPACK model continues to serve as a foundational framework for guiding design education in the context of AIGC. This section provides the theoretical grounding for examining teachers’ perceptions, instructional strategies, and content-related cognition, thereby directly informing RQ2 by clarifying the influence of technological and pedagogical knowledge on AIGC integration in higher design education.

2.4. Research on the Application of the Technology Acceptance Model (TAM) in the Educational Field

The TAM, proposed by Davis in 1986, is a classical theoretical framework for explaining users’ willingness to adopt new technologies [25]. Its core constructs—Perceived Usefulness (PU) and Perceived Ease of Use (PEOU)—have been widely used in educational technology research to predict technology adoption behaviors [26].

For instance, Gong et al. (2004) and Cheung and Vogel (2013) empirically confirmed the significant influence of PU and PEOU on learners’ acceptance of digital learning systems [27,28].

However, recent studies (Lee et al., 2003; Sharp, 2007) have criticized TAM for lacking explanatory depth in complex behavioral contexts, calling for cross-theoretical integrations to improve its robustness in emerging settings such as AIGC [29,30].

In the context of generative AI, tools like ChatGPT and Midjourney exhibit high autonomy, creative unpredictability, and ethical ambiguity. These unique features introduce new challenges for traditional TAM applications, requiring researchers to consider cognitive burden, emotional involvement, and ethical concerns.

Existing studies suggest that while, PU and PEOU remain core predictors of acceptance, learners’ emotional regulation (e.g., anxiety, confidence) and ethical perception (e.g., fairness, responsibility) can significantly moderate the formation of behavioral intention, especially in AI-rich educational environments.

Therefore, this study extends the TAM framework by integrating “emotional regulation” and “ethical perception” as key moderating variables, aiming to better explain students’ behavioral intentions when using AIGC technologies in learning environments.

This provides theoretical support for RQ2, which explores how students’ perceptions of usefulness and ease of use interact with emotional and ethical dimensions to shape their sustainable engagement with AI-powered tools.

2.5. Sustainability-Oriented Educational Innovation Practices

Sustainability has become not only a global educational concern, but also a strategic pillar in AIGC-driven curriculum transformation. In the context of this study, sustainability is regarded as a key lens for evaluating the long-term pedagogical value and ethical implications of intelligent design education systems. Within the domain of design education, sustainability-oriented teaching practices emphasize the integration of long-term impact and environmental adaptability into curriculum design [31]. These practices advocate for educational transformation that fosters design professionals equipped with both innovative capabilities and ecological responsibility.

The recent literature emphasizes that sustainability should be embedded not only in content delivery, but also in the design of instructional systems and evaluation tools. Vitrac et al. (2023) [32], through the “FitNESS” project, designed a packaging curriculum grounded in synthetic materials chemistry and environmental regulations [32]. Their research provides a new paradigm for food packaging design education while enhancing professional competency levels. Chang and Chuang (2021) [33], using a heritage temple as a case study, demonstrated how integrating design thinking and community participation could foster cultural sustainability. Such examples underscore the value of interdisciplinary learning environments in building students’ sustainability competencies [33].

However, most existing research focuses on sustainability content and outcome metrics, lacking integrative theoretical models that combine sustainability with emerging AI tools and behavioral acceptance frameworks. Few studies offer empirical strategies for quantitatively integrating sustainability variables—such as responsibility perception, ethical reasoning, and long-term engagement—into design education models enhanced by generative AI technologies.

To address this gap, the current study builds a sustainability-oriented evaluation model for AIGC-based curricula, using sustainability perception as a mediating variable in technology acceptance and behavioral intention formation. This theoretical approach not only allows for the measurement of educational outcomes, but also the systemic sustainability impact of AIGC in curriculum innovation.

This section supports RQ3 by providing the sustainability and curriculum design rationale for integrating AIGC into higher design education. It outlines both the theoretical underpinnings and operational frameworks necessary to evaluate how generative AI tools contribute to cultivating sustainable design literacy and ethical creative practices among students.

3. Research Methodology

3.1. Overview of Research Methods

This study follows a quantitative-dominant, qualitative-supplementary mixed-methods strategy and employs the following three research approaches:

• Curriculum content analysis: By analyzing design courses from 10 universities in China, this study extracted the integration patterns of AIGC, including introduction methods, instructional role positioning, student task types, and assessment criteria. This provides foundational insights into the educational ecosystem and organizational logic of AIGC-supported instruction.
• Structured questionnaire survey: TAM- and TPACK-based structured questionnaires were separately designed for teachers and students. The surveys include variables such as PU, PEOU, Behavioral Intention (BI), technology anxiety, TPACK competence, and curriculum adaptability [34]. Data were collected using a 5-point Likert scale and were used for subsequent model construction and regression analysis.
• Semi-structured interviews: In-depth interviews were conducted with selected representative teachers and students to explore their perceptions, feedback, and concerns regarding AIGC-based teaching practices. The interview protocol covered topics such as instructional adaptability, tool selection, feedback mechanisms, and ethical considerations [35]. Thematic analysis was applied to interpret the results (see Figure 1).

Through the complementary use and cross-validation of the three research methods, this study aimed to uncover the integration logic of AIGC in university-level design education across three dimensions: instructional structure, teacher–student behavior, and educational philosophy. Accordingly, this study proposes scalable pathways for pedagogical reform. The research design strives to balance theoretical depth with empirical breadth, addressing the pressing demands of current educational technology development while offering scientifically grounded insights for the future integration of generative artificial intelligence and higher education.

3.2. Course and Sample Description

This study selected 10 representative universities from mainland China, including both comprehensive universities and art institutions. The focus is on design-related courses offered between 2022 and 2024 that incorporated AIGC technologies. These courses span various disciplines, including visual communication design, commercial illustration, digital media art, and new media art. A course case archive was established for the 10 universities, and five analytical dimensions were extracted: instructional objectives, methods of tool integration, student task design, teaching strategies, and evaluation mechanisms. These dimensions laid the foundation for classifying integration models and analyzing their characteristics [36]. The course selection followed the criteria below:

• The course explicitly incorporated generative AI tools such as Midjourney, Wenxin Yiyan (ERNIE Bot), Stable Diffusion, or ChatGPT into its instructional content;
• The course featured a complete structure, including components such as lectures, project-based creation, and final evaluations or presentations;
• The availability of course materials was high, with at least three of the following types accessible: syllabus, course structure documentation, course schedule, or screenshots from online teaching platforms.

3.3. Questionnaire Design and Implementation

To gain a deeper understanding of teachers’ and students’ acceptance and behavioral intentions regarding AIGC-integrated teaching, two structured questionnaires were designed, one targeting university design instructors and the other aimed at undergraduate students majoring in design [37]. The questionnaire was developed based on validated constructs from the TAM and the TPACK framework (Davis, 1989; Mishra & Koehler, 2006) [21,25]. Items were adapted to reflect the context of AIGC tools in design education. The development process included expert reviews to ensure content validity. This approach aligns with recent studies employing similar methodologies to assess perceptions and competencies related to generative AI in educational settings (Chan & Hu, 2023; Saúde et al., 2024) [38,39], and it covered the following dimensions:

• The teacher questionnaire included self-assessment of TPACK, AIGC teaching experience, PU, PEOU, BI, and perceived instructional challenges.
• The student questionnaire included AIGC tool usage experience, PU, PEOU, Creative Flow and Cognitive Shift, Individual Expression, and Perceived Technological Dependence.
• Both questionnaires employed a 5-point Likert scale (1 = strongly disagree, 5 = strongly agree) to ensure data comparability and compatibility with statistical analysis methods.

The structure of the questionnaires and the relationship between variables are illustrated in Figure 2, with sample items presented in

Table 1. Comparison table of teacher and student questionnaire dimensions and sample items.

Group	Dimension Name	Theoretical Framework	Sample Item	Literature Source
Teacher	Technological Knowledge	TPACK	I am able to operate AIGC tools such as Midjourney, DALL·E independently.	Davis, 1989 [25];
	Pedagogical Knowledge	TPACK	I can adjust my AIGC teaching strategies based on different student profiles.
	Content Knowledge	TPACK	I understand the core professional knowledge required for my design topics.
	TPACK Integrated Competency	TPACK	I can integrate AIGC tools with course content to enhance teaching outcomes.
Teacher	Perceived Usefulness	TAM	I believe AIGC can effectively improve students’ design quality in class.	Mishra and Koehler, 2006 [21];
	Perceived Ease of Use	TAM	I think AIGC tools are easy and intuitive to operate.
	Behavioral Intention	TAM	I plan to use AIGC tools to support teaching in more courses.
Student	Perceived Usefulness	TAM	I think AIGC tools help me complete design assignments more efficiently.	Mishra and Koehler, 2006 [21];
	Perceived Ease of Use	TAM	I think AIGC tools are user-friendly and easy to learn.
Student	Creative Flow and Cognitive Shift	Exploratory Variable	AIGC helps me more easily overcome creative bottlenecks.	Exploratory item developed by authors
	Individual Expression	Exploratory Variable	AIGC does not compromise the expression of my personal style.
	Perceived Technological Dependence	Exploratory Variable	I worry that long-term reliance on AIGC may weaken my independent thinking.

.

Using a combination of snowball sampling and targeted invitations, the surveys were distributed and collected anonymously via the “Wenjuanxing” online platform. A total of 52 valid teacher responses and 219 valid student responses were obtained, ensuring sufficient sample diversity and representativeness.

To further ensure the reliability and construct validity of the questionnaire instruments, this study conducted reliability and sampling adequacy analyses using SPSS 27.0.

Table 2. Questionnaire constructs, item distribution, and reliability/validity indicators for teachers and students.

Group	Construct	Items	Cronbach’s Alpha	KMO	Bartlett’s Test (p)
Teacher	Technological Knowledge	Q1, Q2	0.87	0.79	χ2 = 371.01, df = 78, p < 0.001
	Content Knowledge	Q10, Q11	0.85
	TPACK Integrated Competency	Q3, Q4	0.78
	Perceived Usefulness	Q5, Q6	0.76
	Perceived Ease of Use	Q7, Q8	0.85
	Behavioral Intention	Q12, Q13	0.81
	Pedagogical Knowledge	Q9	single item
Student	Perceived Usefulness	Q2, Q3, Q10	0.82	0.89	χ2 = 2130.99 df = 78, p < 0.001
	Perceived Ease of Use	Q4, Q5, Q12	0.71
	Creative Flow and Cognitive Shift	Q6,	single item
	Individual Expression	Q1, Q8, Q11, Q13	0.79
	Perceived Technological Dependence	Q7	single item

summarizes the item distribution, Cronbach’s Alpha coefficients, and the results of the sampling adequacy tests (KMO and Bartlett’s Test) for both the teacher and student questionnaires across different constructs. For multi-item constructs, Cronbach’s Alpha values ranged from 0.76 to 0.87 in the teacher group and from 0.71 to 0.82 in the student group, indicating acceptable to high internal consistency. Constructs measured by single items—such as Pedagogical Knowledge (Teacher), Creative Flow and Cognitive Shift (Student), and Perceived Technological Dependence (Student)—were not subject to internal consistency analysis, but were retained due to their conceptual importance and empirical relevance in the context of AIGC-integrated design education.

The Kaiser–Meyer–Olkin (KMO) values were 0.79 for the teacher group and 0.89 for the student group. Bartlett’s Tests of Sphericity were statistically significant in both groups (teacher: χ² = 371.01, df = 78, p < 0.001; student: χ² = 2130.99, df = 78, p < 0.001). These statistical indicators confirm the data quality and support the validity of the subsequent quantitative analysis.

3.4. Semi-Structured Interview Procedures

To complement the quantitative data and capture nuanced teaching experiences and cognitive differences, this study conducted semi-structured interviews with selected teachers and students, following the principle of theoretical saturation. Based on the principle of theoretical saturation, this study sampled 10 teachers and 10 students, representing different teaching roles and experience levels.

The teacher interviews were designed around five key themes: value judgments regarding AIGC in teaching, operational methods, resource constraints, adaptation of TPACK-related competencies and future curriculum outlook (see

Table 3. Teacher interview outline.

Topic	Teachers’ Experiences and Perceptions of AIGC Tool Integration in Higher Education Design Courses
TK	Are you familiar with or have you used AIGC tools (e.g., Midjourney, DALL·E, ChatGPT, etc.)?
	In your teaching or design projects, do you actively guide students in using these tools? How frequently are they used?
	In your opinion, how do AIGC tools affect students’ “visual expression” or “ideation generation”?
PK	Have you ever designed specific teaching activities or sessions focused on AIGC-related content? What were the outcomes?
	In the process of teaching with AIGC, have you encountered any difficulties in instructional organization or pedagogical strategy?
	Have you observed changes in students’ behaviors or engagement methods in class as a result of AIGC integration?
CK	Have you guided students to think critically about AIGC-related ethical issues, such as copyright or originality?
	Do you think AIGC challenges the traditional boundaries between “design as authorship” and “design as content”?
	What do you think is the most enhanced learning goal of AIGC-integrated courses? (e.g., conceptual innovation, cultural value, technical adaptability, etc.)
General Questions	How do you envision the future educational value and development prospects of AIGC in higher education design curricula?
	If an AIGC-integrated course module is to be developed in the future, what forms of academic or institutional support would you need as a teacher?

). The student interviews focused on the following: AIGC usage experiences, PU and PEOU, creative generation experiences, and suggestions for course improvement (see

Table 4. Student interview outline.

Topic	Student Experiences and Perceptions of AIGC Tool Integration in Higher Education Design Courses
Basic Exposure and Tool Usage	Have you heard of or used any AIGC-related tools (e.g., Midjourney, DALL·E, ChatGPT, etc.)?
	In which course(s) did you first come into contact with these tools? Under what circumstances?
	How frequently do you use AIGC tools? Do you use them proactively or only when required by the instructor?
Learning Experience and Tool Perception	Do you think AIGC tools help with your design ideation? If yes, in what specific ways?
	Have you encountered any difficulties or technical barriers while using these tools? What were the main issues?
	Compared to traditional design processes, what changes have you noticed in your workflow when using AIGC tools?
Teaching Guidance and Classroom Feedback	Did your instructor clearly explain how to use AIGC tools and help you understand the learning goals of the course?
	Do you think using AIGC tools made the class more interesting or more confusing and chaotic? Why?
	If AIGC tools were not used, how do you think your design assignments or final project outcomes would have been different?
Cognitive Change and Value Perception	While using AIGC to assist in design, have you considered issues related to originality, copyright, or fairness?
	Do you believe the outputs generated by AIGC tools can fully represent your design intentions?
	Has the use of AIGC tools changed your understanding of “design” or your identity as a “designer” or “creator”?
Course Expectations and Suggestions	Do you hope that AIGC tools will continue to be used in future design courses? Why or why not?
	If a specialized AIGC design course were developed, what kinds of learning content would you hope to see included?

). All interviews employed open-ended questions to encourage participants to share their thoughts freely based on their personal teaching and learning experiences. Each interview lasted approximately 35 min and was recorded and transcribed for analysis. The interview data were analyzed using thematic analysis and were coded according to the TPACK analytical framework, yielding categorized findings and observed tendencies regarding AIGC integration from both teacher and student perspectives.

4. Data Analysis

4.1. Analysis of Course Content

To systematically uncover the integration pathways of AIGC in design education curricula, this study constructed a five-dimensional analytical framework grounded in the TPACK model, design thinking principles, and a project-based learning (PBL) orientation. The framework encompasses five core dimensions: instructional objectives, methods of tool integration, task structure, instructional organization, and assessment mechanisms. The categories within each dimension and their corresponding coding schemes are presented in

Table 5. Five-dimensional coding framework for AIGC course instructional analysis.

Analytical Dimension	Subcategories and Examples	Coding Label
Instructional Objective Type	T1 = Skill Acquisition, T2 = Tool Literacy, T3 = Creative Thinking-Oriented, T4 = Other	T1, T2, T3, T4
Tool Integration Method	I1 = Auxiliary Integration, I2 = Modular Embedding, I3 = Full-Process Embedding	I1, I2, I3
Task Structure Type	S1 = Single-Task Type, S2 = Staged-Task Type, S3 = Comprehensive Project-Based Type	S1, S2, S3
Instructional Organization	M1 = Traditional Lecture, M2 = Project-Based, M3 = Theory-Practice Hybrid	M1, M2, M3
Assessment Mechanism	E1 = Outcome-Based Assessment, E2 = Process + Outcome-Based Assessment, E3 = Multi-Agent Evaluation	E1, E2, E3

. This analytical framework not only helps standardize semantic interpretations of course content, but also provides a structural foundation for the subsequent classification of integration models and pathway modeling.

In terms of analytical methodology, this study builds upon the TPACK theoretical framework while integrating principles from PBL and design thinking pedagogy to construct a five-dimensional course analysis model [40]. The framework encompasses five core dimensions: instructional objectives, tool integration strategies, task structure, instructional organization, and evaluation mechanisms. Using semantic extraction and categorical coding, course syllabi, project briefs, classroom materials, and student outputs were systematically collected and analyzed. Based on this process, a coding matrix was developed to summarize the primary integration patterns and developmental trends of AIGC-enabled courses in higher education. To illustrate the practical application of the five-dimensional framework, this study systematically coded ten representative AIGC design courses across Chinese universities. The results, categorized by each of the five dimensions, are presented in

Table 6. Encoded structure of ten AIGC courses across five pedagogical dimensions.

Name	Instructional Objective	Tool Integration	Task Structure	Instructional Organization	Assessment Mechanism	Coding Structure
University A	T3	I1	S1	M3	E2	T3-I1-S1-M3-E2
University B	T3	I2	S1	M2	E3	T3-I2-S1-M2-E3
University C	T4	I2	S2	M2	E1	T4-I2-S2-M2-E1
University D	T3	I2	S1	M3	E2	T3-I2-S1-M3-E2
University E	T3	I2	S2	M2	E2	T3-I2-S2-M2-E2
University F	T3	I2	S2	M2	E1	T3-I2-S2-M2-E1
University G	T4	I3	S1	M2	E1	T4-I3-S1-M2-E1
University H	T3	I3	S2	M3	E2	T3-I3-S2-M3-E2
University I	T4	I1	S2	M1	E1	T4-I1-S2-M1-E1
University J	T3	I1	S1	M2	E1	T3-I1-S1-M1-E1

. To enhance clarity and brevity, abbreviated university names are used in the table.

The analysis results indicate that, in terms of instructional objectives, most courses prioritize the development of students’ creative thinking rather than focusing solely on technical skills or tool proficiency. Approximately 70% of the courses explicitly identify “stimulating creative generation” or “enhancing freedom of visual expression” as core teaching goals. This reflects a common understanding among universities that AIGC tools serve as catalysts for cognitive training and expressive transformation in design, rather than merely functioning as production tools [41].

Regarding the mode of tool integration, the most prevalent approach is the “module-embedded” model, accounting for roughly half of all courses. In such cases, AIGC tools are incorporated into intermediate tasks or specific stages—such as sketch generation or image style experimentation—forming a partial linkage structure of “tool–task–output.” About 30% of the courses adopt a “supportive integration” strategy, in which instructors demonstrate AIGC tools or present case-based explorations to highlight their potential for inspiring creativity and expanding design language. Notably, only one course implemented a “fully embedded” model where AIGC tools were integrated throughout the entire process—from research and ideation to creation and presentation—signaling an early shift toward a “tool–thinking co-construction” paradigm in teaching strategies.

In terms of task structure, most courses adopt either a “single-task” or “stage-based task” approach. These designs emphasize the use of AIGC tools in specific phases to complete clearly defined subtasks, such as composition, style simulation, or text-to-image transformation. Only one course adopted a “comprehensive project-based” structure, attempting to construct a full human–AI co-creation process spanning from problem definition and ideation to generation, review, and refinement. This model places higher demands on instructors’ TPACK competencies, teaching workflow coordination, and assessment strategies, and remains a relatively rare yet promising innovation.

Regarding instructional organization, the “project-based” and “workshop-style” formats are particularly prominent. About 90% of the courses employ a hybrid model combining workshops with lectures, emphasizing process-oriented collaboration and iterative feedback. Some courses further integrate online platforms, supporting multimodal and cross-platform teaching practices.

At the same time, significant differences are observed in the evaluation mechanisms. While most courses still rely primarily on final outcomes as the basis for assessment, some universities have started to experiment with “process-and-outcome combined” and “multi-stakeholder evaluation” systems. These approaches incorporate feedback from instructors, students, and peers, embedding AIGC-supported instruction within a continuous cycle of “learning–generating–evaluating.”

To visually represent the integration distribution across the five key instructional dimensions, the coding results of the ten AIGC-integrated design courses were statistically analyzed and visualized. As shown in Figure 3, the results highlight notable differences in integration types related to tool usage, task structure, teaching methodology, and evaluation strategies.

As shown in the figure, “module-embedded” integration accounts for the highest proportion in the dimension of tool usage, indicating that AIGC tools are currently employed primarily as mid-course interventions within instructional tasks. The frequent adoption of “project-based/workshop-style” organizational models and “outcome-oriented” evaluation mechanisms reflects the widespread establishment of practice-driven instructional structures across universities. In contrast, models such as “fully embedded integration” and “multi-source evaluation” remain in the exploratory phase, suggesting that there is still room for deeper integration.

Based on the comparative coding across all dimensions, this study further identifies three representative types of AIGC-integrated instructional models: the Concept-Supporting Model, Tool-Driven Model, and Project-Embedded Model.

The Concept-Supporting Model focuses on stimulating students’ design thinking, where AIGC tools serve a primarily inspirational role in the early stages of the course. The Tool-Driven Model emphasizes technological empowerment, with AIGC playing a central role in content generation. The Project-Embedded Model pursues full alignment among instructional objectives, teaching processes, and assessment mechanisms, demonstrating a high level of systemic integration in instructional design.

These three types differ significantly in terms of depth of tool integration, instructional orientation, teacher role positioning, and student participation models. The structural logic of the three integration pathways is illustrated in Figure 4. The coexistence and transitional nature of these models reflect the current diversity in AIGC instructional practices and point toward a potential shift from “tool utilization” to “curricular reconstruction” in future integration strategies.

4.2. Questionnaire Data Analysis

4.2.1. Student-Side Data Analysis

In the PU dimension, students rated AIGC’s ability to enhance work quality Q2(M = 4.20) and improve design efficiency Q3(M = 4.22) highly. Similarly, PEOU scores showed that students found the tools accessible Q4(M = 3.92) and easy to operate Q5(M = 3.88). Creative stimulation received the highest score Q6(M = 4.28), reflecting strong engagement with the technology.

However, some concerns were identified in the ethical and expressive dimensions. Scores related to weakened personal expression Q7(M = 3.56), ethical awareness Q9(M = 3.32), and overreliance concerns Q10(M = 3.31) showed higher variability, with SD values exceeding 0.9.

Students expressed strong continued usage motivation: exploratory intent Q11(M = 4.15), willingness to use in future courses Q12(M = 4.18), and long-term integration attitudes Q13(M = 4.21).

To visualize these patterns, seven key items (Q2, Q3, Q6, Q7, Q9, Q11, Q13) were selected to construct a radar chart, as shown in Figure 5. This provides an overview of students’ multi-dimensional perceptions toward AIGC technologies.

Figure 5 illustrates that students rated creative stimulation, efficiency enhancement, and continued intention (Q6, Q3, Q13) most highly. In contrast, lower scores in Q7 and Q9 reflect lingering concerns about personal expression and ethical implications. These insights suggest that future AIGC-integrated design instruction should focus not only on enhancing usability, but also on reinforcing students’ critical thinking and ethical literacy.

To streamline data presentation and an avoid item-by-item analysis, seven key questionnaire items were selected to represent core dimensions such as perceived usefulness, creativity, ethical concern, and behavioral intention. Their mean scores and standard deviations are shown in

Table 7. Descriptive statistics of key student evaluation variables.

Variable Code	Variable Description	Mean Score	Standard Deviation
Q2	Perceived Usefulness—Design Enhancement	4.20	0.80
Q3	Perceived Usefulness—Efficiency Improvement	4.22	0.78
Q6	Degree of Creative Thinking Activation	4.28	0.74
Q7	Perceived Difficulty in Controlling AIGC Output	3.56	0.86
Q9	Ethical Concern Awareness	3.32	1.03
Q11	Individual Expression—Willingness to Explore AIGC Technologies	4.15	0.83
Q13	Individual Expression—Intention of Continued AIGC Use	4.21	0.77

.

As shown in Figure 5, the highest scores were observed for Q6(M = 4.28), Q3(M = 4.22), and Q13(M = 4.21), corresponding to creative stimulation, efficiency improvement, and intention for continued use, respectively. This indicates that students generally acknowledge the practical value of AIGC technologies in design learning.

Conversely, lower scores on Q7 and Q9 suggest potential concerns regarding personal expression and ethical awareness. These results highlight the importance of future instructional strategies that not only enhance students’ technical control over AIGC tools, but also cultivate ethical literacy and critical thinking. Such efforts are essential for advancing a more creative, responsible, and sustainable approach to design education.

4.2.2. Teacher-Side Data Analysis

The reliability analysis showed a Cronbach’s Alpha of 0.88, a KMO value of 0.80, and a statistically significant result in Bartlett’s test of sphericity (p < 0.001), indicating strong suitability for factor analysis. The Principal Component Analysis (PCA) extracted three factors, with a cumulative explained variance of 67.98%, supporting the construct validity of the TPACK framework.

In terms of TK, PK, and CK, the results reflect teachers’ generally positive evaluations across all three dimensions. Teachers demonstrated adequate proficiency in understanding and operating AIGC tools, but expressed relatively lower confidence in guiding students, highlighting areas for professional development in TK.

In PK, most teachers recognized AIGC’s potential to enhance creativity and instructional efficiency, while also acknowledging challenges such as output unpredictability and technological disparity among students. The CK dimension received the highest scores, with particular emphasis on the importance of ethical instruction and institutional support for long-term curricular integration.

To better illustrate these findings,

Table 8. Descriptive statistics of mean scores and standard deviations for TPACK dimensions in the teacher questionnaire.

Dimension	Item (Sample)	Mean Score	Standard Deviation
TK	Guide students in AIGC use	3.52	0.79
PK	AIGC enhances creativity	4.15	0.67
CK	Institutional training needed	4.31	0.65

summarizes the descriptive statistics for representative items in each TPACK dimension.

Figure 6 further visualizes the average performance across the TK, PK, and CK dimensions, indicating that CK scores were consistently higher than TK and PK, suggesting stronger teacher confidence in content-related integration.

To further examine the structural relationships among the three TPACK dimensions,

Table 9. Pearson correlation matrix of the three TPACK dimensions.

	TK	PK	CK
TK	1.00	0.69 ** [0.51, 0.81]	0.53 ** [0.30, 0.70]
PK	0.67 ** [0.51, 0.81]	1.00	0.42 ** [0.16, 0.62]
CK	0.53 ** [0.30, 0.70]	0.416 ** [0.16, 0.62]	1.00

Note: Pearson correlation coefficients are reported with 95% confidence intervals in brackets. Double asterisks indicate significance at ** p < 0.01.

shows that there are significant positive correlations among the three dimensions.

The correlation analysis revealed significant positive relationships among the three dimensions, especially between TK and PK (r = 0.69, p < 0.01) and between TK and CK (r = 0.53, p < 0.01), providing empirical support for the internal consistency of the TPACK framework in this context.

In summary, teachers exhibit strong readiness in CK and moderate performance in TK and PK, suggesting a need for targeted training in AIGC tool application and pedagogical strategy design. Strengthening technical workshops and fostering collaborative teaching models could facilitate a shift from isolated tool use to strategic curriculum integration, thereby supporting the sustainable transformation of higher design education.

4.3. Path Modeling and Validation on the Teacher Side

To further explore the interrelationships among the three TPACK dimensions in the context of AIGC-integrated design instruction, this study constructed a multiple linear regression model with CK as the dependent variable and TK and PK as independent variables. The aim was to verify whether teachers’ technical competencies and instructional strategies could effectively predict their level of recognition regarding AIGC-related course content, ethical considerations, and pedagogical integration.

4.3.1. Model Construction and Regression Analysis

Based on the previously defined TPACK dimensional framework, this study calculated the average scores of teachers across three dimensions: TK, PK, and CK. These averages were used as the core independent variables in the regression analysis. The regression model is specified as follows:

$${\mathrm{C}\mathrm{K}}_i={\beta }_0+{\beta }_1\cdot {\mathrm{T}\mathrm{K}}_i+{\beta }_2\cdot {\mathrm{P}\mathrm{K}}_i+{\epsilon }_i$$

Here, ${\mathrm{C}\mathrm{K}}_i$ denotes the content knowledge score of the $i$ teacher (where $i$ = 52). ${\mathrm{T}\mathrm{K}}_i$ and ${\mathrm{P}\mathrm{K}}_i$ represent the average scores of the teacher in the dimensions of Technological Knowledge and Pedagogical Knowledge, respectively. ${\beta }_0$ is the constant term, ${\beta }_1$ and ${\beta }_2$ are the regression coefficients, and ${\epsilon }_i$ is the error term. The regression analysis was conducted using Python 3.11 and the statsmodels statistical library, with Ordinary Least Squares (OLS) estimation. The results are shown in

Table 10. Results of the OLS regression analysis.

Variable	Regression Coefficient	Standard Error	p-Value
Constant	3.11	0.37	<0.001 ***
TK	0.26	0.09	0.007 **
PK	0.07	0.12	0.56 (n.s.)

Note: *** p < 0.001, ** p < 0.01, n.s. = not significant.

.

The model’s coefficient of determination was R² = 0.32, and the adjusted R² = 0.30, indicating that the regression model explains approximately 29.80% of the variance in the Content Knowledge outcome dimension.

4.3.2. Analysis Results and Interpretation

According to the regression results, TK showed a statistically significant effect on CK, with a regression coefficient of β = 0.26. This suggests that, the more familiar teachers are with AIGC-related technologies, the higher their cognitive competence in integrating course content, addressing ethical issues, and applying AIGC in instructional contexts.

Although PK also showed a positive coefficient (β = 0.07), it did not reach statistical significance (p = 0.56), indicating that, in this sample, pedagogical strategies had a relatively weak direct impact on CK. It is possible that PK exerts influence through mediating or interactive pathways rather than as a direct predictor.

This outcome reflects a practical condition in the early stages of AIGC curriculum integration. While many instructors possess well-developed PK from traditional teaching, their existing strategies may not yet be fully adapted to or transformed by the rapid evolution of AIGC technologies. As a result, the direct effect of PK on CK remains limited. In contrast, AIGC technologies are highly operational and generative by nature, requiring teachers to engage in hands-on experimentation and deep interaction in order to internalize and transform such experiences into meaningful curriculum content and ethical awareness.

To visualize the influencing pathways among TPACK variables on the teacher side, a structural path model was developed, as shown in Figure 7.

Therefore, TK is not merely a measure of operational familiarity with tools, but also serves as an internal driving force for content recognition, ethical orientation, and curriculum design. This finding offers a new interpretation of the TPACK framework within the context of AIGC-integrated teaching.

Specifically, in the early phase of rapid technological emergence, the influence of TK on CK appears to be more direct, while PK may function as a mediator or moderator. Further validation of these relationships could be pursued using more advanced structural equation modeling (SEM) approaches.

4.4. Summary of Interview Results and Course Feedback Analysis

4.4.1. Teacher Perspective: Experiences and Challenges in Integration Practice

Overall, most teachers expressed a positive attitude toward the instructional potential of AIGC technologies, and yet they also reported various challenges in terms of pedagogical structure, content value guidance, and ethical considerations.

At the level of TK, most participants demonstrated a basic understanding and operational proficiency with AIGC tools, particularly platforms like Midjourney, DALL·E, and ChatGPT. Some educators have incorporated these tools into classroom demonstrations and case discussions with moderate fluency. Teachers generally agreed that AIGC tools effectively stimulate students’ creativity in both visual expression and conceptual ideation. In project-based courses especially, AIGC helps students rapidly construct initial visual styles and semantic frameworks. One interviewee noted:

“After being exposed to AIGC, students became less constrained by technical limitations during the sketching phase—they are now more imaginative and able to realize their ideas.”

However, challenges were more prominent in the dimension of PK. On the one hand, many teachers admitted that neither they nor their institutions have developed a clear instructional framework for AIGC integration. Most practices rely on ad hoc “plug-in” tool applications within existing curricula, lacking systematic instructional structures or evaluative alignment. On the other hand, during classroom implementation, teachers reported difficulties in maintaining focus on design fundamentals. Due to AIGC’s strong automatic generation features, students may become overly reliant on output results in the short term, thereby weakening training in essential competencies such as design process thinking, conceptual reasoning, and esthetic judgment. Some teachers also observed a shortage of diversified task models and reflective mechanisms tailored to AIGC-based instruction, resulting in a recurring pattern of “having tools but lacking methodology.”

In the dimension of CK, teachers demonstrated heightened sensitivity to how AIGC redefines the boundaries of design education. On the one hand, many acknowledged that AIGC disrupts traditional notions of “design” by shifting creation from “human-directed tool use” to “co-creation with algorithms”, prompting them to reconsider the definitions of originality, process, and expression in design. On the other hand, teachers raised concerns regarding ethics, copyright, and bias in algorithmic training data, and some proactively incorporated these topics into classroom discussions. These include guiding students to reflect on content ownership and esthetic agency in the AI age. One teacher remarked:

“Students are initially excited, but I have to help them realize that this is not just a shortcut—it may also reshape their understanding of what it means to be a creator.”

Overall, teachers have shown a relatively rapid adaptation to AIGC technologies, demonstrating strong technical learning abilities and a clear willingness to integrate new tools. However, a systematic gap persists in pedagogical design and content transformation. While teachers expressed an open yet cautious attitude toward the ethical implications of AIGC, its impact on the role of design education, and the evolving mindset of students, they also acknowledged a pressing need for curricular support.

These insights corroborate the findings from the earlier questionnaire analysis, where TK significantly predicted CK, whereas PK did not show a significant direct effect. Together, the data and interview responses reveal a transitional pattern in AIGC-based instruction that can be characterized as follows: Technology-driven → Methodologically lagging → Content-reconstructing.

To achieve the deeper integration of AIGC into design curricula, future faculty development initiatives must strengthen the triadic capacity of “Technology–Pedagogy–Content” and construct an instructional ecosystem that is responsive to emerging technological paradigms.

4.4.2. Student Perspective: Acceptance and Perception of AIGC-Integrated Instruction

In terms of tool exposure and usage habits, most interviewees reported having initial contact with AIGC tools via courses, social media, or project assignments—particularly with Midjourney and ChatGPT, which were frequently cited. Many students first used these tools during early ideation or sketching phases of design projects. While some were encouraged by instructors to explore these tools, others were self-motivated by curiosity. Nonetheless, AIGC was mostly used as a situational aid, without forming part of a consistent creative workflow.

Regarding learning experiences and tool perception, student feedback revealed a dual attitude. On the one hand, most students acknowledged the convenience of AIGC in stimulating creativity and generating visual concepts, noting that it helped them “quickly visualize abstract ideas,” especially by generating diverse stylistic and contextual variations in a short time. On the other hand, several students encountered significant barriers, such as difficulty formulating effective prompts, unpredictability in output, and an inability to clearly articulate personal design intent. One student noted: “I gave it a very detailed description, but the output didn’t match my vision—it left me more confused.” Moreover, the black-box nature of these tools was frequently mentioned as a source of cognitive dissonance: “AI seems powerful, but I don’t really understand why it produces these results.”

In classroom organization and instructional support, students provided divergent opinions. Some appreciated the basic tool demonstrations provided by instructors, but noted a lack of guidance regarding design goals, creative direction, and evaluation criteria, leading to a state of “disorganized exploration.” Others critiqued the course for introducing AI without emphasizing the design process or conceptual logic, stating: “It generates images fast, but not what I intended—it doesn’t convey the meaning I hoped for.” Compared to traditional hand-drawing or software-based modeling, one recurring reflection was that “AI makes us feel more like editors than creators.”

At the cognitive level, students expressed concerns about authorship and originality. Several questioned the ownership of AIGC-generated work and whether such content could truly be attributed to them. Ethical questions were raised, such as “Should every AI-generated result be signed?” and “Do I have the authority to interpret what it created?” The introduction of AIGC has also redefined students’ understanding of the “designer” identity, from someone who creates manually to someone who curates and inputs ideas, prompting deeper reflection on their creative role.

In summary, students generally demonstrated a high level of acceptance toward AIGC in design education, especially valuing its utility for efficiency and idea generation. However, concerns remain regarding precision in creative expression, control over outputs, authorship, and instructional clarity. These findings align closely with the questionnaire results, particularly on items Q6(creativity enhancement), Q7(lack of output control), and Q9(ethical awareness), confirming that current AIGC integration resides between the “technological enthusiasm phase” and the “cognitive adaptation phase.” To achieve meaningful integration, course design must enhance the link between tool usage and design logic and foster students’ sense of authorship responsibility and ethical awareness throughout the creative process.

5. Research Discussion and Instructional Strategy Recommendations

5.1. Summary of Key Research Findings

By integrating course content analysis, questionnaire surveys, path modeling, and semi-structured interviews, this study offers a comprehensive picture of the current status, challenges, and underlying mechanisms of AIGC integration in higher education design courses. Framed by the TAM and the TPACK framework, the study identifies key patterns in technical acceptance, pedagogical adaptation, and curricular cognition among both teachers and students.

First, the course content analysis reveals that the application of AIGC tools in current design curricula primarily occurs in phases such as image generation, concept development, and creative expression. These applications are largely characterized by task-based introduction and demonstrative operations, rather than systematic modular integration [42]. This reflects the exploratory stage of AIGC use in education, where curricular design remains centered on tool functionality instead of instructional goals. Consequently, most courses reflect a form of technological embedding rather than true pedagogical integration.

Second, in the student survey results, variables such as PU, Creative Stimulation (Q6), and Learning Efficiency (Q3) received significantly higher scores than other dimensions. These results indicate broad student approval of AIGC’s effectiveness in enhancing design efficiency and creative output. However, lower average scores and greater variance were found for dimensions such as Output Controllability (Q7) and Ethical Awareness (Q9), suggesting divergent levels of understanding and value judgment among students. Importantly, students demonstrated strong long-term willingness to use AIGC (Q13) and curiosity to explore its potential (Q11), which indicates that AIGC has a promising foundation for sustainable educational integration.

Third, in the path modeling analysis of teacher data, TK significantly predicted CK (β = 0.26, p < 0.01), while PK showed no significant impact on CK (β = 0.07, p = 0.56). These finding highlights that, at this early stage of AIGC integration, teachers’ capacity to align content with emerging technologies is primarily driven by their level of technical understanding, rather than established pedagogical strategies. This aligns with the structural logic of the TPACK framework, which posits that, in the early phases of technological transformation, TK serves as the core driver of instructional integration.

Lastly, in the interview analysis, both teachers and students expressed a general recognition of AIGC’s pedagogical potential. Teachers agreed that AIGC facilitates creativity in visual expression and conceptual ideation, yet also identified several key challenges: ambiguous tool integration strategies, student over-reliance on generated content, and difficulty aligning outputs with instructional objectives. Students echoed similar sentiments: while acknowledging AIGC’s convenience in visual production and stylistic exploration, they also raised concerns over uncontrollable outputs, disconnects between tool and design intent, and unclear authorship in AI-assisted creations.

In summary, this study identifies a common structural pattern in current AIGC-integrated design instruction in Chinese higher education: strong technical potential, weak application pathways, and underdeveloped instructional structures. On the teacher hand, the key contradiction lies in “technology-driven instruction—lagging pedagogical development—insufficient content integration.” On the student side, learners are navigating a transitional phase between initial excitement and cognitive adaptation, wherein instructional guidance remains inadequate. These multi-dimensional findings offer empirical support and logical grounding for future efforts in curriculum reform, teacher training frameworks, and instructional resource development.

5.2. Discussion of the Integration Mechanism Based on the TPACK Framework

Based on the findings from the path modeling analysis grounded in the TPACK framework, this study identifies a structural imbalance in the current integration of AIGC into design education. TPACK emphasizes the organic interplay between TK, PK, and CK as a prerequisite for effective teaching innovation. However, this study reveals that, while a positive link has been established between TK and CK, the mediating role of PK remains underdeveloped. This is reflected in the insignificant path coefficient from PK to CK (β = 0.12, p > 0.05), indicating that pedagogical scaffolding has not yet effectively transformed content knowledge within AIGC contexts.

Specifically, TK functions as the primary driver for AIGC integration, directly influencing whether teachers can meaningfully apply AIGC tools in content design and course construction. The standardized path coefficient from TK to CK was β = 0.37 (p < 0.01), supporting the assumption that technology familiarity strongly affects how instructors incorporate generative tools into curricula. Survey and interview data both suggest that teachers who are proficient in tools such as Midjourney and ChatGPT are more likely to explore their application in aligning technology with course content. However, the lack of corresponding pedagogical training results in fragmented teaching activities, weak task structures, and an absence of stable instructional workflows and assessment mechanisms.

Another indicator of the underperformance of PK is the limited pedagogical scaffolding between technology and content. According to the TPACK model, PK is not merely a set of instructional methods, but rather the critical node that activates the integrative potential of TK and CK. In this study, although teachers acknowledge the educational value of AIGC, few have established a systematic instructional cycle encompassing task design, evaluation mechanisms, and learning feedback. This is further evidenced by low scores in the pedagogical strategy item group in the TPACK survey (mean = 3.19), suggesting a lack of operational teaching routines. This structural gap may explain why PK fails to significantly predict CK in the path model.

From the student perspective, the lack of pedagogical guidance similarly impacts their learning experiences and cognitive construction. When AIGC is presented merely as a technical demonstration in class—without clear instructional scaffolding or staged learning objectives—students tend to equate generation with final results, neglecting the underlying design thinking and expressive logic. Interview responses also support this finding, with students noting a lack of constructive feedback and limited understanding of the rationale behind AIGC tool use.

This diverges from the TPACK model’s goal of tripartite synergy in technology, pedagogy, and content. To ensure the effective implementation of the TPACK framework in AIGC-integrated design education, it is essential to enhance the mediating role of PK through systematic teacher training, continuous updates in pedagogical theory, and the integration of cross-platform instructional resources. These insights confirm the need for pedagogical reinforcement to balance the current “tech-content” bias and to realize the full potential of AIGC for sustainable educational transformation. Establishing such an integrative mechanism will facilitate the transition of higher design education from isolated tool adoption toward a comprehensive curricular transformation, thereby advancing a more sustainable, theory-informed, and learner-centered model of instructional innovation.

5.3. Key Challenges and Adaptation Strategies in AIGC-Integrated Instruction

Based on empirical data and qualitative interview results, this study identifies three major challenges currently confronting AIGC-integrated instructional practice: (1) cognitive imbalance arising from varying levels of technological adaptation; (2) organizational gaps due to a lack of pedagogical scaffolding; and (3) value conflicts triggered by unclear ethical awareness.

First, inconsistent levels of technological literacy pose a primary barrier to effective implementation. Although most students demonstrate a strong willingness to engage with AIGC tools, they exhibit significant disparities in operational proficiency, output control, and personal expression. Survey data show that over 38% of students reported difficulties in understanding how to optimize prompts or interpret output logic (mean score for “confidence in using AIGC tools” = 2.96). Difficulties in prompt engineering or the opaque logic of generative outputs often lead to reverse cognitive responses, where users report feeling “increasingly confused the more they use it.” In interviews, several students described AIGC tools as “black boxes,” underscoring the need for cognitive scaffolding in instructional design. On the instructor side, while most educators have mastered basic tool operations, many still lack confidence in guiding students effectively, with 42% of teacher respondents rating themselves “not confident” in supporting student tool use, exacerbating both “learning gaps” and “instructional anxiety” within the classroom.

Second, the absence of structured pedagogical frameworks has limited AIGC implementation to demonstrative or supplemental use. Faculty commonly report that AIGC remains a temporary “inserted unit” within existing curricula, lacking cohesive integration with core instructional goals or sustained scaffolding. This is supported by curriculum content analysis results, which show that 70% of AIGC-related instructional content appears only as isolated activities or single sessions rather than being embedded into continuous project-based learning. This weakens the long-term pedagogical value of such tools and undermines students’ sense of direction, ultimately affecting both learning outcomes and motivation.

Third, ambiguous understandings of authorship and ethics are generating value tensions within design education. Interview data reveal that students are often confused about issues such as authorship ownership, algorithmic control, and definitions of originality. Some students explicitly questioned, “If the AI made it, do I still get credit as the designer?” Meanwhile, instructors face uncertainty regarding whether and how to address ethical and copyright concerns, and often lack adequate resources or instructional frameworks to do so. Among surveyed teachers, only 21% indicated having included ethical topics in their AIGC-related teaching. If unaddressed, these tensions could challenge the core educational values of human-centered creativity-driven design pedagogy.

To achieve a more coherent integration of AIGC into instructional systems, it is essential to address these challenges at their structural roots. Transitioning from tool adoption to instructional fusion requires not only the enhancement of technical competencies, but also the development of robust pedagogical systems and the reconfiguration of instructional goals. Three adaptive strategies are proposed to support this transition: (1) implement differentiated training programs to close the gap in technological confidence among students and instructors, as the current data highlight severe imbalances in skill self-assessment between the two groups; (2) establish task-driven and process-oriented instructional structures to align generative tools with pedagogical objectives, addressing the fragmentation identified in curriculum analysis; and (3) embed ethical reasoning and originality awareness into curricular cores, thereby guiding students toward a rational understanding of creative agency and technological mediation.

These steps are crucial for moving from “technology acceptance” toward “knowledge construction” in AIGC-enhanced education, thereby aligning pedagogical innovation with the sustainable development of higher design education.

5.4. Strategic Recommendations for Sustainable Integration of AIGC in Design Education

Based on the findings and challenges identified in this study, the integration of AIGC in higher education is not merely a matter of technological adaptation, but also a critical test of whether educational systems can sustain long-term development in the context of digital transformation [43]. This section proposes a set of strategic recommendations aimed at constructing a future-oriented, sustainable model of design education.

Establishing a Continuous Professional Development System for Educators. Higher education institutions should develop professional training systems centered on “technological innovation and instructional design,” equipping instructors with both operational proficiency in AIGC tools and the ability to transform these tools into meaningful pedagogical practices [44]. Support mechanisms should include regular workshops, pedagogical innovation seminars, and cross-departmental resource-sharing platforms to promote both vertical advancement in technological literacy and horizontal collaboration among faculty.

Structuring and Iteratively Updating Course Modules. Curricular frameworks should recognize AIGC competency as a component of foundational design literacy. Modular course designs should reflect progressive levels of difficulty, enabling students to advance from basic tool awareness to higher-order skills in design transformation. Feedback loops between modules should be implemented to support continuous course improvement and content renewal.

Enhancing Students’ Sustainable Learning and Media Literacy. Instruction should extend beyond technical training to foster students’ capacity for self-directed learning, critical thinking, and ethical reasoning within open-ended technological environments. Accordingly, course content should incorporate topics such as algorithmic logic, creative authorship, and AI ethics, guiding learners to develop enduring learning skills and adaptive capabilities in evolving digital contexts.

Building an Open, Shareable Educational Resource and Assessment Ecosystem. An open resource infrastructure should be developed to support AIGC-integrated education, including repositories of teaching cases, generative image databases, and exemplary student projects. Evaluation frameworks should also be diversified, incorporating formative assessment, peer review, and traceable generation chains for student work to facilitate the sustainable accumulation and dissemination of educational outputs.

Aligning Instructional Innovation with the UN Sustainable Development Goals (SDGs). AIGC-based education should be aligned with key SDGs, particularly SDG 4 (Quality Education) and SDG 9 (Industry, Innovation, and Infrastructure), to shift the focus from “technological instrumentalism” to “value-oriented innovation.” This approach fosters a new generation of designers equipped with both technical expertise and a strong sense of social responsibility [45,46].

Collectively, these strategies offer practical pathways for the deep integration of AIGC in design education and respond to the broader imperative for structural transformation and sustainable evolution in higher education under the dual pressures of post-pandemic adaptation and AI advancement.

5.5. Research Limitations and Future Directions

Despite the systematic exploration of AIGC integration pathways in higher education design curricula through a mixed-methods approach, this study has several limitations that should be acknowledged.

First, the number of teacher participants was relatively limited, with only 52 valid questionnaire responses collected. This sample size may constrain the robustness of the path effects within the structural model. Future research is encouraged to expand the teacher sample base while maintaining consistency in survey dimensions, incorporating respondents from a broader range of regions, institutional types, and teaching backgrounds to enhance statistical power and generalizability.

Second, the structural path model employed in this study was primarily based on linear relationships derived from the TAM and the TPACK framework. Further studies could consider adopting SEM or multi-group path analysis to explore more complex interactions and mediating mechanisms among variables.

Third, although this study proposed several integration strategies oriented toward sustainable teaching, their practical effectiveness has not yet been empirically validated. Future investigations may implement instructional intervention experiments or longitudinal tracking studies to quantitatively assess the real-world outcomes and behavioral impacts of the proposed strategies.

Finally, the rapid advancement of AIGC technologies poses new challenges to teaching models. Future research may explore more intelligent, human-centered, and sustainable instructional systems in design education by incorporating cutting-edge AI tools such as large language models and multimodal interactive platforms.

6. Conclusions

6.1. Research Conclusions

This study focuses on the sustainable integration of AIGC into design education, systematically examining the current applications, cognitive differences, integration pathways, and long-term development potential of AIGC technologies in higher education design curricula. By employing a multi-method approach, including curriculum content analysis, questionnaire surveys, path modeling, and qualitative interviews, this research constructs an analytical framework based on the TAM and the TPACK model. From both teacher and student perspectives, it reveals the actual dynamics and core issues associated with AIGC implementation in design education.

The findings indicate that AIGC technologies hold strong potential in enhancing creative ideation and improving design efficiency [47]. Students generally affirm the instrumental value of these tools, particularly in concept development and visual generation tasks, and express a high willingness to adopt them. However, issues such as output unpredictability and varied ethical awareness introduce cognitive uncertainty [48]. On the teacher hand, there is a dual trend: while TK positively facilitates CK, the role of PK in supporting integration pathways remains underdeveloped. Path modeling results further confirm that TK significantly predicts CK, whereas PK does not exhibit statistically significant influence.

The interview results support the following findings: although instructors generally acknowledge the pedagogical potential of AIGC, they face difficulties in curriculum planning, instructional methods, and evaluation mechanisms due to a lack of systemic design and theoretical guidance. Students also expressed high expectations for clearer course structure and stronger authorship in creative output. This reflects the current developmental phase of AIGC teaching practice as one of “early-stage technological adoption, delayed pedagogical adaptation, and pending content reconstruction.”

Based on these insights, this study proposes five strategic recommendations aimed at fostering sustainable AIGC-integrated instruction. These include building teacher training systems for technical capacity, optimizing modular curriculum structures, enhancing students’ media literacy and ethical awareness, establishing open access educational resource platforms, and aligning pedagogical reform with the UN SDGs. Together, these strategies address the current bottlenecks in AIGC pedagogical integration and offer a theoretical and practical framework for the resilient transformation of education systems in the era of artificial intelligence.

6.2. Research Limitations

Despite the systematic exploration of theoretical frameworks and empirical methods, this study has several limitations. First, the sample is primarily drawn from a subset of Chinese universities, limiting the representativeness in terms of geographic diversity and institutional types. As a result, the findings may not fully capture the conditions of design education programs at varying levels or in different international contexts. Second, the regression path model is based on cross-sectional data, which restricts the ability to observe the dynamic evolution of AIGC integration over time. Additionally, the model does not include mediating variables or employ SEM to further investigate the deeper interrelationships among constructs. Third, the curriculum content analysis and interview data mainly focus on visual communication and illustration design, without extending to other subfields such as product design or interaction design. This disciplinary concentration may limit the generalizability of the study’s conclusions across the broader spectrum of design education.

6.3. Future Research Directions

Future studies may expand along several key dimensions. First, it is recommended that future studies broaden the sampling scope to include a wider range of universities across different regions, types, and academic levels. Comparative analyses—both horizontal and longitudinal—can enhance the external validity of the findings. Second, future research could employ advanced statistical techniques, such as SEM, mediation analysis, and multi-group comparisons, to explore the underlying mechanisms among TPACK dimensions and reveal more nuanced integration pathways.

In addition, subsequent investigations may extend the study of AIGC integration to a broader range of curriculum domains, such as interaction design, game design, and XR-based immersive learning environments. Longitudinal teaching experiments should also be considered to evaluate the practical effectiveness and sustainability of instructional strategies over time. Lastly, increased attention should be given to cross-cultural perspectives. Examining the similarities and differences in AIGC-integrated teaching across diverse educational systems and cultural contexts will contribute valuable theoretical insights into the global transformation of higher education [49].

In summary, as a transformative force in education, the sustainable integration of AIGC in design education necessitates a coordinated effort anchored in teacher capacity building, centered on pedagogical restructuring, and driven by student cognitive engagement. Such an approach will foster the evolution of higher design education toward a more open, intelligent, and sustainable future.