GenAI Learning for Game Design: Both Prior Self-Transcendent Pursuit and Material Desire Contribute to a Positive Experience

Abstract

This study explores factors influencing positive experiences with generative AI (GenAI) in a learning game design context. Using a sample of 26 master’s-level students in a course on AI’s societal aspects, this study examines the impact of (1) prior knowledge and attitudes toward technology and learning, and (2) personal value orientations. Results indicated that both students’ self-transcendent goals and desire for material benefits have positive correlations with collaborative, cognitive, and affective outcomes. However, self-transcendent goals are a stronger predictor, as determined by stepwise regression analysis. Attitudes toward technology were positively associated with cognitive and affective outcomes during the first week, though this association did not persist into the second week. Most other attitudinal variables were not associated with collaborative or cognitive outcomes but were linked to negative affect. These findings suggest that students’ personal values correlate more strongly with the collaborative, cognitive, and affective aspects of using GenAI for educational game design than their attitudinal attributes. This result may indicate that the design experience neutralizes the effect of earlier attitudes towards technology, with major influences deriving from personal value orientations. If these findings are borne out, this study has implications for the utility of current educational efforts to change students’ attitudes towards technology, especially those that encourage more women to study STEM topics. Thus, it may be that, rather than pro-technology instruction, a focus on value orientations would be a more effective way to encourage diverse students to participate in STEM programs.

1. Introduction

Game-based learning is an innovative educational approach that leverages computer games as tools for instruction, integrating elements of interactivity and engagement with instructional goals across various software applications. This method has gained significant traction as it provides opportunities for active learning, supports diverse learning styles, and enhances teaching, assessment, and learner evaluation processes [1]. According to the Game-Based Learning Market report, the global game-based learning market is projected to reach approximately USD 29.7 billion by 2026, reflecting the increasing recognition of its value and effectiveness in education [1].

The educational benefits of game-based learning are multi-faceted, extending beyond mere engagement. Research has shown that incorporating games into educational contexts can enhance learning outcomes, such as problem-solving skills, motivation, and critical thinking [2]. Additionally, designing learning games can provide added value by offering extra learning opportunities. “Learn by design” is a pedagogical approach that emphasizes active, hands-on learning, where students construct knowledge through designing artifacts, such as educational games, that communicate complex concepts to others [3,4]. Designing a learning game that teaches others encourages students to critically analyze the knowledge they are imparting, anticipate learners’ needs, and structure information in accessible ways, all of which deepen their cognitive engagement with the material [5]. The process also enhances students’ metacognitive awareness, as they reflect on how to convey information meaningfully, thereby promoting a mastery-oriented learning environment [6]. Moreover, designing learning games also enhances collaboration, creativity, and intrinsic motivation, as students feel empowered by creating meaningful learning experiences for others [7].

Collaborative learning in game-based contexts adds an additional layer of complexity and richness to the educational experience. Numerous studies highlight the benefits of collaborative and cooperative learning activities, demonstrating their impact on cognitive and affective outcomes, such as increased engagement, positive attitudes, and persistence [8,9,10,11]. Empirical studies have found that small group collaborative learning will enhance students’ enjoyment, interest, learning, and task performance [12,13]. Additionally, it can foster group cohesiveness, improve self-esteem, and reduce error rates through mutual support and shared understanding [14,15]. These outcomes are especially relevant in learning game design settings, where shared goals and teamwork can lead to a richer learning experience.

The field itself is facing a revolution due to generative AI (GenAI) technology [16] and the possibility for students to self-actualize and exercise new powers of creativity and learning. Given this revolution, we aim to contribute to an understanding of the factors that will be influencing the educational enterprise and the students’ role within it.

GenAI technologies offer novel possibilities for creating and customizing educational games. Integrating GenAI into learning game design allows for dynamic adaptation of content to individual learner needs, thereby facilitating personalized learning experiences [17,18]. GenAI is uniquely positioned to support learning game design through its ability to process vast amounts of data, generate novel content, and adapt dynamically to learners’ needs [19,20]. These capabilities align closely with the demands of designing educational games, which require creativity, problem-solving, and collaboration among students. By leveraging GenAI, students can access tools that facilitate brainstorming, automate technical tasks, and simulate real-world scenarios, enabling them to focus more on critical thinking and design innovation. Moreover, studies also suggest that integrating GenAI in educational settings can influence learners’ perceptions and engagement with the material, for example by providing corrective feedback, generating practice exercises, or creating an extended study plan to further enhance the learning experience [21]. The prior literature investigates how these unique affordances of GenAI influence students’ positive experiences during the game-design process. However, the unique dynamics of GenAI in collaborative game design still remain underexplored, particularly in terms of their impact on collaborative, cognitive, and affective outcomes. This study uniquely positions GenAI as both the subject and the tool of inquiry. The learning program is designed to allow students to first learn to effectively use GenAI tools to design learning games, to absorb AI knowledge through the game creation process itself, and ultimately to be able to teach each other with games of their own design.

This research takes an exploratory approach to uncover how specific attitudinal and personal factors—such as students’ goals, personality traits, and collaborative orientations—predict positive experiences in GenAI-driven learning game design. By understanding these predictors, this study aims to contribute to the growing body of research on game-based learning and GenAI, offering insights into optimizing educational games for enhanced learner outcomes and satisfaction.

This study makes three key contributions to the field of educational technology and game-based learning. First, it demonstrates the importance of personal value orientations in predicting positive collaborative, cognitive, and affective outcomes in GenAI-enhanced learning. Second, it identifies a short-term change in the effect of students’ prior knowledge, attitudes toward technology and learning, and personal value orientations on learning outcomes. Third, it provides practical insights for educators and policymakers by emphasizing the potential of value-oriented interventions to encourage diverse participation in STEM fields, particularly among underrepresented groups.

2. Literature Review and Theoretical Perspectives

Among the rich palette of theoretical perspective that bear on our area of interest, we have identified three major lenses or perspectives that can provide a conceptual framework for understanding how GenAI influences learning processes: constructivism, socio-cognitive theory, and self-determination theory. Due to space limitations, we can only offer the briefest overview of each before identifying the perspective that we believe will offer the most insight.

Constructivist theory posits that learning is an active, constructive process where learners build new knowledge based on their prior experiences and interactions with their environment [22]. In the context of GenAI and game design pedagogy, constructivism emphasizes the role of students as co-creators of their educational experiences. Game design activities powered by GenAI offer students opportunities to experiment, prototype, and iterate in a dynamic, learner-centered environment. Socio-cognitive theory (SCT) [23] emphasizes the interplay of individual cognition, social interaction, and environmental factors in learning. From this perspective, GenAI’s role in game design pedagogy lies in enhancing collaborative processes and fostering shared mental models among learners. AI tools can act as mediators in collaborative environments, facilitating communication and mutual understanding. Self-determination theory (SDT) [24] emphasizes the importance of autonomy, competence, and relatedness in fostering intrinsic motivation and positive emotional experiences. In game design pedagogy, GenAI can serve as a tool to enhance students’ intrinsic motivation by reducing barriers to creative expression and increasing feelings of competence.

This brief sketch suggests that there are attractive features of each perspective. However, we have decided to draw on SCT because it offers a particularly compelling lens for examining the role of GenAI in collaborative learning environments. This is because it highlights the importance of shared mental models and collective problem-solving. Within SCT, several themes emerge that are particularly relevant to GenAI and game design pedagogy. First, socio-cognitive theory underscores the importance of collaborative processes in learning, positing that effective teamwork relies on a shared understanding of goals, roles, and tasks among group members. GenAI tools can facilitate these processes by synthesizing group inputs, visualizing complex ideas, and ensuring equitable participation. Second, socio-cognitive theory emphasizes the role of environmental scaffolds—resources or tools that support learners in achieving higher levels of cognitive functioning. GenAI serves as such a scaffold, providing students with access to creative and analytical resources that enhance their ability to design and refine educational games. Finally, socio-cognitive theory highlights the bidirectional influence of individual cognition and group dynamics, suggesting that tools like GenAI can shape not only individual learning but also the collective efficacy and cohesion of teams.

While prior studies have applied SCT to traditional collaborative learning [25], its application to GenAI-driven environments remains sparse. This gap is significant, as GenAI tools introduce unique affordances (e.g., real-time content generation) that may alter group problem-solving and communication patterns. By adopting SCT, this study bridges the theoretical divide between GenAI’s technical capabilities and the social processes essential for effective collaboration.

From this lens, we derive our variables of interest: transcendental goals (such as contributing to societal well-being or creating meaningful educational experiences) [26] and material goals (such as achieving high performance or producing a polished final product) [27]. These variables align with socio-cognitive theory’s emphasis on the interplay between individual motivations and group processes. Examining these goals through a SCT lens allows us to investigate how they influence group dynamics, shared mental models, and the effectiveness of GenAI tools in collaborative contexts.

2.1. Prior Knowledge and Attitudes Towards Technology and Learning: 5 Predictors

In selecting the five variables—attitudes toward technology, confidence in using computers, attitudes toward using GenAI for teaching and learning, attitudes toward collaborative group projects, and perceived prior knowledge of machine learning—we aimed to capture a broad spectrum of student dispositions that could shape their experience and effectiveness in using GenAI for learning game design.

Students’ general attitudes towards technology are likely to affect their experience in integrating GenAI into learning game design. Research shows that a positive orientation towards technology can foster openness and adaptability when using novel tools in educational settings [28] and workplace [29]. While studies like [28,29,30] highlight a correlation between tech-positive attitudes and tool adoption, conflicting evidence suggests this relationship is context-dependent. For instance, ref. [31] found that, despite their enthusiasm for technology, educators’ proficiency in using socio-communication technologies was limited due to factors such as a lack of domain-specific skills and hands-on experience. Furthermore, longitudinal studies [32] reveal that initial tech optimism may wane when learners encounter technical barriers, underscoring the need to distinguish between transient attitudes and sustained engagement.

Confidence in using computers might correlate with students’ engagement and success in technology-mediated learning game design experiences. Bandura’s social cognitive theory [23] suggests that higher confidence in specific skills leads to increased resilience, persistence, and creativity when facing challenges. Therefore, students with greater confidence in computer usage may have a more positive experience overall when working on AI-driven educational games. Therefore, this attribute is a critical predictor of how well students can adapt and engage with the project.

We also took into account students’ attitudes towards using GenAI for teaching and learning since the project requires students to apply GenAI tools in designing games that teach other students AI knowledge. Those with a more favorable view of AI’s role in education are likely to perceive this experience as meaningful and aligned with educational innovation, potentially resulting in more engagement, collaboration, and satisfaction (see review at [33,34]). However, recent critiques [35] reveal that enthusiasm for AI often masks practical challenges, such as accuracy, algorithmic bias or transparency issues. For example, learners who initially embrace AI-generated content may later express frustration when tools produce inconsistent, indirect, or vague outputs [32,36]. This suggests that attitudes toward GenAI are not static but evolve with experiential exposure—a nuance overlooked in cross-sectional studies. As we hope to show with our exploratory study that this important gap benefits from disambiguation.

Positive attitudes towards collaborative projects are known to encourage teamwork, communication, and the development of shared mental models [25]. These factors contribute to improved group dynamics, efficient role distribution, and effective resolution of conflicts, all of which are crucial in learning game design [25]. Consequently, students with a positive orientation towards group work are more likely to enjoy and benefit from collaborative aspects, making it an essential attitudinal predictor of their positive experience. However, the success of collaborative learning hinges on nuanced factors. Ref. [37] emphasize that positive attitudes toward group work improve effective teamwork, but this relationship falters in poorly structured tasks. For instance, studies found that role ambiguity created technostress [38] for users that disrupted human-AI collaboration and hinder their adoption [39]. This aligns with socio-cognitive critiques: while attitudes toward collaboration reflect social interaction, their impact is contingent on task design and shared goals. Few studies explore how GenAI’s generative capabilities (e.g., real-time content creation) alter collaboration learning outcomes. This is one of the areas we hope to contribute to with our analysis.

Perceived prior knowledge of machine learning can influence students’ confidence and approach to learning game design that teaches AI concepts. Self-assessment of prior knowledge has been linked to increased engagement and comfort when working with complex subject matter, as familiarity with foundational concepts can reduce cognitive overload and facilitate deeper learning [40]. Therefore, in this study, students who feel they have a baseline understanding of machine learning may experience fewer barriers and more motivation to apply and expand this knowledge through game design.

2.2. Personal Value Orientations: Self-Transcendent Goals and Material Benefits

In selecting self-transcendent goals and material benefits as focal personal value orientations, we aimed to capture two contrasting orientations that reflect both intrinsic and extrinsic factors influencing students’ experience in GenAI learning game design. These attributes address both students’ desire to impact others positively and their pragmatic career aspirations, both of which are relevant in educational contexts where technical and social skills converge.

Self-transcendent goals reflect an individual’s inclination to contribute positively to the well-being and learning of others, aligning with intrinsic motivations that enhance collaborative and altruistic endeavors [41,42]. Research has consistently demonstrated that self-transcendence is linked to personal fulfillment, community involvement, and motivation in various domains, including learning [43], mental health [44], and spirituality [45].

Given the above theoretical base, individuals with higher levels of transcendental goals may experience greater satisfaction when working in AI-related fields. This is particularly true in areas where the expansion of individual potential is a well-promoted cornerstone [46]. Now shifting our theoretical development, we call to the reader’s attention the following comments. Specifically, in the context of GenAI learning game design, students with strong self-transcendent goals may experience greater satisfaction in creating games that educate peers about AI, as their motivations extend beyond self-gain to a sense of community impact and collective learning. These students are likely to exhibit higher resilience, creativity, and empathy, all of which are crucial for collaborative game design and for effectively communicating complex AI concepts to others. Thus, self-transcendent goals may positively shape students’ positive experience with GenAI for learning purposes, as their intrinsic desire to contribute aligns with the educational purpose of the project.

The pursuit of material benefits represents a pragmatic, career-oriented perspective on learning and adopting AI skills, which can also affect students’ experiences in designing AI-based learning games. Material aspirations, such as the desire for financial success, recognition, and career advancement, can serve as extrinsic motivators that drive students to excel in acquiring and applying AI competencies [47]. In this study, students who view their participation in the project as a stepping stone to achieving financial or professional status may experience an enhanced motivation to develop practical skills and demonstrate proficiency, which can lead to a positive engagement with the project’s technical and educational aspects. Moreover, since GenAI is a field with high potential for lucrative career opportunities, students who are focused on material benefits might view the project as an avenue for professional growth and skill acquisition, which can increase their commitment and sense of purpose in the learning process. However, while extrinsic motivations might not always lead to a deep, intrinsic engagement with GenAI for education, they can nonetheless foster a productive learning environment where students are motivated by the perceived value and potential rewards of their efforts in GenAI game design.

2.3. Collaborative Outcomes: Group Collaboration and Shared Mental Models

Collaboration within educational settings has long been associated with enhanced learning outcomes [15,48]. In learning game design, collaboration plays an essential role in enabling students to exchange knowledge, refine ideas, and integrate diverse perspectives that are central to both game development and an accurate representation of AI concepts [48]. Effective collaboration can lead to a deeper understanding of AI principles, as students learn to navigate problem-solving and creative processes as a collective. This aligns with literature suggesting that perceived collaboration not only supports cognitive engagement but also fosters a supportive learning environment where students feel empowered to share and build upon each other’s ideas, ultimately enhancing both the quality of the final product and the students’ satisfaction and motivation [48,49].

A second collaborative outcome that this study explored is the perceived group shared mental model. A shared mental model refers to a common understanding among team members regarding key aspects of their task, including goals, procedures, and each other’s roles and strengths [50]. Research on group dynamics has shown that a strong shared mental model improves coordination and reduces misunderstandings, allowing students to work more efficiently and make decisions that are better aligned with their collective objectives [50,51]. Thus, exploring how AI-facilitated learning game design influences students’ shared mental models contributes insights into how students internalize and collectively interpret AI knowledge, thereby enhancing the learning experience for both the designers and their future learners.

2.4. Cognitive Outcomes: Flow and Creative Problem-Solving

Cognitive flow refers to a state of complete immersion and focus, often associated with high levels of intrinsic motivation and the optimal blending of challenge and skill [52]. Achieving flow is beneficial for learning as it fosters a positive emotional connection to the task, often resulting in greater persistence and satisfaction [53]. Examining cognitive flow in this setting is, therefore, crucial to understanding how students experience the design process and how deeply they engage with AI concepts in a meaningful, enjoyable way.

Creative problem-solving is the second cognitive outcome in this study. Engaging students in creative problem-solving during game design allows them to apply AI concepts in novel and meaningful ways, thereby deepening their understanding of these concepts while enhancing their cognitive capacity for future problem-solving tasks. Investigating creative problem-solving in this study context sheds light on how students approach and overcome design challenges, revealing the cognitive processes through which they assimilate and operationalize AI knowledge.

2.5. Affective Outcomes: Intrinsic Motivation and General Negative Affect

Intrinsic motivation reflects a deep, self-driven interest in the task [54]. In designing a learning game, intrinsic motivation can drive students to explore, experiment, and invest more effort into creating an effective and engaging product. Intrinsically motivated students are more likely to experience the task as valuable and stimulating, resulting in higher cognitive engagement and persistence in the face of challenges [54]. By exploring students’ intrinsic motivation, this study addresses how the design process itself can serve as a source of enjoyment and inspiration, potentially leading to more effective learning outcomes and a stronger, long-term interest in AI and educational technology.

In addition, assessing students’ general negative affect during the game design process offers valuable insights into potential obstacles and emotional challenges they may face. Negative affect can include feelings such as frustration, boredom, or stress [55]. Monitoring negative affect might be relevant in AI-based game design, where students encounter new and challenging concepts that may increase stress or reduce motivation if not appropriately managed.

Therefore, this study raised the following hypotheses, which are stated in general form in the interests of parsimony:

H1. 

Students’ positive prior knowledge and attitudes towards technology and learning predict a positive collaborative, cognitive, and affective experience in GenAI learning game design.

H2. 

Students’ positive personal value orientation predicts a positive collaborative, cognitive, and affective experience in GenAI learning game design.

3. Methods

3.1. Research Design and Participants

This study employs an exploratory experimental design within the graduate-level course “AI in Society” to examine whether students can develop a learning game using GenAI that educates others on foundational AI concepts. The design involves 12 student groups, each consisting of 2–3 members. Each group is assigned one of two topics for their game: machine learning basics or AI in K-12 education. The game design process spans two weeks. Participants are asked to complete a self-report survey at the end of each week’s class.

Convenience sampling was used to recruit 26 young adults (23 females and 3 males) between the age of 21 to 39 from a large northeastern university. All participants provided informed consent prior to the study. The research received approval from the university’s Institutional Review Board to ensure adherence to ethical standards.

3.2. Measures

Attitudes towards technology. Attitudes towards technology were measured with 9 five-point Likert style items adapted from [56,57] (scale running from 1, “strongly disagree”, to 5, “strongly agree”). Example items were “I am comfortable learning new technology” and “today it is essential that everyone knows about computers”. These nine items proved to be highly reliable (α = 0.91, M = 3.73, SD = 0.68).

Confidence using computers. Confidence using computers was measured with 5 five-point Likert style items adapted from [58] (scale running from 1, “strongly disagree”, to 5, “strongly agree”). Example items were “generally, I feel OK when faced with a new problem on the computer” and “I’m sure I could learn a computer language”. Negatively worded items were reverse coded. These five items proved to be highly reliable (α = 0.83, M = 3.57, SD = 0.72).

Attitudes towards using AI for teaching and learning. Attitudes towards using AI for teaching and learning was measured with 4 five-point Likert style items adapted from [59] (scale running from 1, “strongly disagree”, to 5, “strongly agree”). Example items were “AI makes learning more interesting” and “I look forward to lessons that require me to use AI”. These four items proved to be highly reliable (α = 0.88, M = 4.31, SD = 0.61).

Attitudes towards collaborative group projects. Attitudes towards collaborative group projects were measured with 3 five-point Likert style items (scale running from 1, “strongly disagree”, to 5, “strongly agree”). Example items were “I enjoy collaborating with other students on group projects” and “working in a group helps me understand the material better”. These three items proved to be highly reliable (α = 0.77, M = 3.53, SD = 0.70).

Perceived prior knowledge on machine learning. Perceived prior knowledge on machine learning was measured with 6 five-point Likert style items (scale running from 1, “strongly disagree”, to 5, “strongly agree”). Example items were “I know a lot about machine learning” and “I know machine learning more than most people”. These six items proved to be highly reliable (α = 0.84, M = 3.21, SD = 0.69).

Self-transcendent goals. Self-transcendent goals were measured with 6 five-point Likert style items adapted from [41] (scale running from 1, “strongly disagree”, to 5, “strongly agree”). Example items were “I wish to use my AI knowledge to serve others” and “I wish to use AI to help people with physical and mental difficulties”. Negatively worded items were reverse coded. These six items proved to be highly reliable (α = 0.97, M = 4.60, SD = 0.47).

Material benefits. Material benefits were measured with 3 five-point Likert style items (scale running from 1, “strongly disagree”, to 5, “strongly agree”). Example items were “I want to make a lot of money using my AI skills” and “I will be valued because of my skills with AI”. These three items proved to be highly reliable (α = 0.81, M = 4.32, SD = 0.65).

Collaboration. Student’s collaboration on the learning game design was measured with 11 five-point Likert style items (scale running from 1, “strongly disagree”, to 5, “strongly agree”). Example items were “collaboration within our group was effective” and “collaboration within our group facilitated a smooth development process”. These eleven items proved to be highly reliable in both Week 1 (α = 0.96, M = 4.32, SD = 0.65) and Week 2 (α = 0.97, M = 4.43, SD = 0.76).

Shared mental model. Shared mental model was measured with 8 five-point Likert style items (scale running from 1, “strongly disagree”, to 5, “strongly agree”). Example items were “members of the team have similar sources for problem-solving resources (e.g., technologies, equipment, and tasks)” and “members of the team have similar standards about the importance of task-related resources”. These eight items proved to be highly reliable in both Week 1 (α = 0.97, M = 4.44, SD = 0.56) and Week 2 (α = 0.96, M = 4.50, SD = 0.61).

Cognitive flow. Student’s cognitive flow was measured with 9 five-point Likert style items adapted from [52,53,60] (scale running from 1, “strongly disagree”, to 5, “strongly agree”). Example items were “I feel just the right amount of challenge” and “my thoughts run fluidly and smoothly”. These nine items proved to be highly reliable in both Week 1 (α = 0.95, M = 4.08, SD = 0.69) and Week 2 (α = 0.97, M = 4.16, SD = 0.79).

Creative problem-solving. Student’s creative problem-solving was measured with 8 five-point Likert style items adapted from [41] (scale running from 1, “strongly disagree”, to 5, “strongly agree”). Example items were “in this game design task, I generate many new ideas” and “in this game design task, I create different solutions for a problem”. These eight items proved to be highly reliable in both Week 1 (α = 0.95, M = 4.34, SD = 0.47) and Week 2 (α = 0.90, M = 4.44, SD = 0.48).

Intrinsic motivation. Student’s intrinsic motivation was measured with 18 five-point Likert style items adapted from [53,55,61,62] (scale running from 1, “strongly disagree”, to 5, “strongly agree”). Example items were “I liked this activity” and “I enjoyed doing this activity very much”. These eighteen items proved to be highly reliable in both Week 1 (α = 0.98, M = 4.46, SD = 0.54) and Week 2 (α = 0.99, M = 4.35, SD = 0.64).

Negative affect. Student’s negative affect was measured with 9 five-point Likert style items adapted from [55] (scale running from 1, “strongly disagree”, to 5, “strongly agree”). Example items were “during the game design, I felt distressed” and “during the game design, I felt scared”. These nine items proved to be highly reliable in both Week 1 (α = 0.96, M = 1.77, SD = 0.81) and Week 2 (α = 0.94, M = 1.60, SD = 0.61).

4. Results

Students’ various experiences during the learning game design are demonstrated on

Table 1. Students’ game design experiences across two weeks.

Variable	Week 1		Week 2		t-Test	Hedges’ g Effect Size
	M	SD	M	SD
Collaboration	4.52	0.47	4.43	0.76	t(43.31) = 0.51 p = 0.610	0.138
Shared mental model	4.44	0.56	4.50	0.61	t(51.61) = −0.35 p = 0.730	−0.093
Cognitive flow	4.08	0.69	4.16	0.79	t(51.03) = −0.40 p = 0.693	−0.106
Creative problem-solving	4.34	0.47	4.44	0.48	t(51.99) = −0.77 p = 0.445	−0.207
Intrinsic motivation	4.46	0.54	4.35	0.64	t(50.72) = 0.65 p = 0.521	0.173
Negative affect	1.77	0.81	1.60	0.61	t(48.44) = 0.85 p = 0.397	0.229

and Figure 1 below. A series independent sample t-test shows that there are not significant differences between the two-week design period.

To test H1, we conducted a Pearson correlation analysis to examine the relationship between students’ prior knowledge and attitudes toward technology and learning and their experiences in game design. To address H2, we conducted a Pearson correlation analysis to assess how students’ personal value orientations correlated with their game design experiences. The results for Week 1 are presented in

Table 2. Correlation between pre-existing attitudinal variables and game design experiences variables in Week 1.

	Collaboration	Shared Mental Model	Cognitive Flow	Creative Problem-Solving	Intrinsic Motivation	Negative Affect
Attitudes towards technology	0.27	0.29	0.47 **	0.38 **	0.42 **	−0.32 *
Confidence using computers	0.26	0.16	0.23	0.13	0.29	−0.51 ***
Attitudes towards using AI for teaching and learning	−0.25	−0.10	0.01	−0.03	0.14	−0.05
Attitudes towards collaborative group projects	0.15	0.21	0.11	−0.01	0.10	−0.43 **
Perceived prior knowledge on machine learning	−0.23	−0.01	0.22	0.14	0.11	−0.40 **
Self-transcendent goals	0.53 ***	0.50 ***	0.41 **	0.51 ***	0.59 ***	−0.64 ***
Material benefits	0.26	0.36 *	0.19	0.35 *	0.42 **	−0.37 *

Note: One-tailed test was deployed on the correlation coefficients, * p ≤ 0.1; ** p ≤ 0.05; *** p ≤ 0.01.

, and those for Week 2 are in

Table 3. Correlation between pre-existing attitudinal variables and game design experiences variables in Week 2.

	Collaboration	Shared Mental Model	Cognitive Flow	Creative Problem-Solving	Intrinsic Motivation	Negative Affect
Attitudes towards technology	0.05	0.14	0.00	−0.01	0.02	−0.31
Confidence using computers	0.18	0.22	−0.09	−0.24	−0.00	−0.20
Attitudes towards using AI for teaching and learning	−0.04	−0.07	−0.02	−0.13	0.20	−0.32 *
Attitudes towards collaborative group projects	0.37 *	0.33 *	0.13	0.02	−0.00	−0.24
Perceived prior knowledge on machine learning	0.15	0.14	0.04	−0.05	0.07	−0.39 **
Self-transcendent goals	0.42 **	0.36 *	0.42 **	0.28	0.3 7 *	−0.33 *
Material benefits	0.32	0.32 *	0.28	0.26	0.23	−0.33 *

Note: One-tailed test was deployed on the correlation coefficients, * p ≤ 0.1; ** p ≤ 0.05.

below.

We further performed stepwise regression analyses to identify which combinations of students’ pre-existing attitudinal variables best predicted their game design experiences. In total, we conducted 12 stepwise regression analyses for the seven game experience variables, with separate analyses for each of the two weeks.

Week 1

Results show that the combination of self-transcendent goals, perceived prior knowledge on machine learning, and attitudes towards technology in general most effectively predict student collaboration on the game design, F(3,20) = 8.586, adjusted r² = 0.497, p < 0.001, Cohen’s f = 0.573. The other attitudinal independent variables did not enter the regression.

Self-transcendent goals most effectively predict students’ shared mental model in the design process, F(1,22) = 8.326, adjusted r² = 0.242, p = 0.009, Cohen’s f = 0.249. The other attitudinal independent variables did not enter the regression.

Attitudes towards technology in general most effectively predict student’s cognitive flow during the game design, F(1,22) = 6.834, adjusted r² = 0.202, p = 0.016, Cohen’s f = 0.206. The other attitudinal independent variables did not enter the regression.

Self-transcendent goals most effectively predict students’ creative problem-solving during the design, F(1,22) = 7.137, adjusted r² = 0.211, p = 0.014, Cohen’s f = 0.216. The other attitudinal independent variables did not enter the regression.

Self-transcendent goals most effectively predict students’ intrinsic motivation during the design, F(1,22) = 12.298, adjusted r² = 0.329, p = 0.002, Cohen’s f = 0.348. The other attitudinal independent variables did not enter the regression.

The combination of self-transcendent goals and confidence using computers most effectively predict students’ negative affect during the design, F(2,21) = 12.805, adjusted r² = 0.507, p < 0.001, Cohen’s f = 0.588. The other attitudinal independent variables did not enter the regression.

Week 2

Self-transcendent goals most effectively predict student collaboration on the game design, F(1,21) = 5.317, adjusted r² = 0.164, p = 0.031, Cohen’s f = 0.166. The other attitudinal independent variables did not enter the regression.

Self-transcendent goals most effectively predict students’ cognitive flow during the game design, F(1,21) = 5.395, adjusted r² = 0.167, p = 0.030, Cohen’s f = 0.169. The other attitudinal independent variables did not enter the regression.

Self-transcendent goals most effectively predict students’ shared mental model in the design process, F(1,21) = 8.326, adjusted r² = 0.242, p = 0.009, Cohen’s f = 0.249. The other attitudinal independent variables did not enter the regression.

Material benefits most effectively predict students’ negative affect during the design, F(1,21) = 4.417, adjusted r² = 0.134, p = 0.048, Cohen’s f = 0.135. The other attitudinal independent variables did not enter the regression.

5. Discussion

5.1. Noble Self-Transcendent Goals Lead to Positive Design Experiences

Results revealed that the higher students’ self-transcendent goals are, the more positive experiences they have during the game design, including better collaboration with teammates, better shared mental model, better cognitive flow, better performance on creative problem-solving, higher intrinsic motivation, and lower negative affect. In Week 1, the combination of self-transcendent goals, perceived prior knowledge on machine learning, and attitudes towards technology in general exert a strong practical impact on students’ collaborative experiences (Cohen’s f = 0.537 > 0.35; Cohen’s f interpretation guidelines see [63]). This highlights the strong practical significance of these predictors in fostering effective teamwork in GenAI learning game design. Self-transcendent goals moderately predict students’ ability to develop a shared mental model (0.15 < Cohen’s f = 0.249 < 0.35) and students’ ability to generate novel and effective solutions during the design process (0.15 < Cohen’s f = 0.216 < 0.35), and strongly predict intrinsic motivation (Cohen’s f = 0.348 = 0.35). This indicates that shared values and altruistic motivations play a meaningful role in aligning group goals and understanding and the importance of altruistic motivations in fostering creativity. Students who are driven by altruistic purposes appear to derive greater enjoyment and fulfillment from the design process. The combination of self-transcendent goals and confidence in using computers significantly reduces students’ negative emotions during the design process (Cohen’s f = 0.588 > 0.35). This suggests that these predictors provide strong emotional resilience in the face of challenges.

In Week 2, self-transcendent goals have a modest impact on collaboration (0.15 < Cohen’s f = 0.166 < 0.35), maintaining cognitive flow (0.15 < Cohen’s f = 0.169 < 0.35), fostering shared mental model (0.15 < Cohen’s f = 0.249 < 0.35). The smaller effect on collaboration compared with Week 1 may reflect a shift as collaboration becomes more influenced by other factors, such as group dynamics or task familiarity. The medium effect size of self-transcendent goals on maintaining cognitive aligns with the idea that intrinsic values continue to drive engagement even as the task progresses. The consistency of medium effect size on fostering shared understanding among team members suggests that such goals are a stable predictor of group alignment.

Self-transcendent goals focus on benefiting others and contributing to the well-being of society rather than pursuing personal achievements or individual success alone. In this context, students aspire to use AI knowledge and skills for the betterment of others and society, such as helping people with physical or mental challenges, designing AI applications for broader benefit, and creating a positive global impact. Therefore, with self-transcendent goals, students are likely to experience these positive outcomes because such goals foster a sense of purpose that aligns with various positive experiences.

The shared purpose may unify students, making them more supportive and effective as a team. Students are also more likely to develop a similar vision or approach to the project. This purpose-driven approach can lead to greater motivation, engagement, and enjoyment, as they view their efforts as having a positive impact beyond personal achievement. Additionally, when students see their projects as ways to contribute to societal well-being or to help others, they are more likely to experience satisfaction and fulfillment, which can lead to more creativity and resilience in tackling design challenges. These values make the process feel rewarding and personally meaningful, fueling more positive and enriching experiences throughout the design journey.

5.2. Desire for Material Benefits Correlated with Positive Design Experiences

In Week 2, material benefits have a modest influence on reducing negative affect (0.02 < Cohen’s f = 0.135 < 0.15). While less impactful than in Week 1, the relationship suggests that students’ desire for material benefits from AI plays a significant role in shaping their experience during the learning game design process. In this context, students with higher aspirations for material benefits demonstrated better collaborative outcomes, notably through improved shared mental models within their groups, as well as creative problem-solving. The association between material benefits and shared mental models may indicate that as students anticipate professional gains from their AI skills, they may place higher value on achieving a cohesive understanding with their teammates, fostering a collaborative atmosphere where ideas and responsibilities are clearly shared and understood. The link between material benefits and creative problem-solving may suggest that students motivated by future career and financial rewards may be more inclined to engage deeply with the design challenge, viewing it as an opportunity to develop and showcase valuable skills. This finding aligns with previous research suggesting that extrinsic motivations can enhance task-oriented engagement and problem-solving abilities in learning contexts that are perceived as directly relevant to career goals [46].

Notably, material benefits were the strongest predictor of reduced negative affect. This suggests that students motivated by extrinsic rewards experienced less frustration, stress, or disinterest, perhaps because the project aligned with their long-term aspirations. While intrinsic motivations are often emphasized in educational research, these findings illustrate that those extrinsic goals, such as material benefits, can also positively influence students’ emotional experiences by reinforcing their sense of value in the work.

5.3. Positive Attitudes Towards Technology Only Foster Positive Design Experiences in Early Stages

The results indicate that students with more positive attitudes towards technology experienced enhanced cognitive and affective outcomes, particularly in the early stages of the learning game design process. In the first week, positive attitudes towards technology were significantly correlated with higher levels of cognitive flow, stronger creative problem-solving, increased intrinsic motivation, and reduced negative affect. However, these correlations and predictive associations did not persist into the second week. In addition, attitudes toward technology in general have a moderate impact on students’ experience of cognitive flow (0.15 < Cohen’s f = 0.206 < 0.35) in Week 1. This suggests that students who are comfortable with technology are better able to immerse themselves in challenging and engaging design tasks.

These findings could be attributed to the novelty and initial excitement of working on an AI-based project. In their first week, students’ positive dispositions towards technology drive their engagement and enjoyment, fostering a productive cognitive environment. However, it is possible that as the design process progresses, students’ experiences become more influenced by practical challenges and task demands rather than initial attitudes. As students encounter deeper complexities or fatigue, their initial enthusiasm may wane, leading to a reduced impact of attitudes towards technology on their experiences. The decline may also reflect the natural progression of task adaptation, where initial excitement gives way to the realities of sustained effort and collaboration, shifting the focus from individual attitudes to group dynamics, task structure, or content knowledge.

5.4. Other Prior Knowledge and Attitudes Towards Technology and Learning Variables Did Not Correlate with Positive Design Experiences

The results indicate that while confidence using computers, attitudes towards using GenAI for teaching and learning, attitudes towards collaborative group projects, and perceived prior knowledge of machine learning did not significantly predict a positive design experience, they were associated with lower negative affect at different stages of the game design process. This may suggest that, although these variables may not directly enhance the collaborative, cognitive, or affective outcomes of the experience, they play a protective role in reducing students’ negative emotions during the project.

In the first week, students who were more confident in using computers and those with positive attitudes toward collaborative group projects reported significantly lower negative affect. This early-stage emotional benefit likely reflects the initial adjustment phase, where technical confidence and collaborative readiness may mitigate anxiety or frustration as students adapt to the project’s requirements. Confidence in computer use, for instance, may alleviate stress associated with navigating AI tools, providing students with a smoother transition into the technical aspects of game design. Similarly, positive attitudes towards group work may ease initial interpersonal challenges, as students with collaborative mindsets may feel more comfortable communicating and coordinating with their teammates. However, as we discussed in the literature review [38,39], as the project progressed and practical challenges intensified, these initial attitudes may have been less influential, which could explain why their correlation with lower negative affect did not persist into the second week.

Interestingly, students’ attitudes towards using GenAI for teaching and learning were not significantly correlated with lower negative affect in the first week but became significant in the second week. This shift suggests that positive attitudes towards GenAI in educational contexts may play a role in sustaining emotional resilience over time. As students move beyond the initial adaptation period and engage more deeply with the purpose and potential impact of their project, those who view GenAI as a valuable educational tool may experience reduced negative affect due to their alignment with the project’s educational goals. As reflected in the literature review [32,36], this finding supports the idea that changes occur as usage progresses. Additionally, our study contributes to the literature by highlighting that these changes may take time to manifest.

Finally, perceived prior knowledge of machine learning consistently correlated with lower negative affect across both weeks. Students who perceive themselves as having a stronger foundation in machine learning may feel more competent and prepared, which can decrease anxiety and frustration throughout the project.

5.5. Design Experiences Did Not Significantly Change over Time

The results indicate that students reported consistently positive experiences throughout the GenAI learning game design project, with stable collaborative, cognitive, and affective outcomes over time. One explanation for these stable experiences could be that the structure of the project effectively supported students’ collaboration and learning. The project likely provided a strong framework that fostered positive interactions and maintained students’ focus. Another possible factor is the alignment between students’ interests and the project’s goals. Many students may have already had an interest in AI and educational technology, which could have contributed to their sustained cognitive and affective engagement.

We have also compared males and females and did not find any significant variations between the two groups. While having an equal number of males and females would have been ideal, we are not making a gender-specific argument or generalization. This obviates an additional problem had we attempted to do so. Moreover, research suggests that gender composition can influence group dynamics and collaboration styles. For instance, studies have found that female-dominated groups often emphasize cooperation, consensus-building, and emotional support, which align closely with the collaborative outcomes measured in our study [64]. These tendencies may have amplified the positive effects of self-transcendent goals on shared mental models and collaboration, as these goals align with communal and prosocial values often associated with female participants. However, this gender imbalance may also have limited the diversity of perspectives and interaction styles, which could have influenced the creativity and problem-solving processes observed in the study. Now covering several threads of discussion, our findings suggest that basic, underlying value orientations, rather than specific short-term attitudes, may have a more substantial and lasting impact on students’ collaborative experiences and emotional engagement in GenAI learning game design. This highlights an important distinction: while short-term, topic-specific attitudes, such as positivity or negativity towards technology, may influence immediate interactions, they appear to diminish over time, especially as students gain familiarity with the technology. In contrast, enduring value orientations—such as self-transcendent goals and material benefit aspirations—are associated with sustained positive outcomes. These findings imply that an individual’s broader outlook and deeply held values shape not only their initial approach to technology-based projects but also their persistence and resilience over time.

6. Implications

The implications of this study are particularly relevant as educators increasingly incorporate GenAI-based learning games as educational tools. Understanding the factors that foster positive experiences in GenAI learning game design—such as students’ value orientations and basic attitudes toward collaboration and technology—can help educators more effectively support students in these activities. For example, teachers could consider fostering self-transcendent goals, alongside career-relevant motivations to improve students’ engagement and reduce negative affect. In addition, educators designing collaborative, AI-driven projects may benefit from considering how gender dynamics shape group interactions. For instance, balancing gender representation within teams or providing explicit support for diverse interaction styles could help create more equitable and effective learning environments. However, given that this is a preliminary study, the findings should be interpreted cautiously, as they do not provide definitive conclusions about the best practices for implementing GenAI educational games. Additional research is needed to verify these insights and further explore how educators can create optimal learning environments that support sustained interest and positive emotional experiences in technology-enhanced education.

The implications of this study could also extend beyond the classroom and suggest potential applications at the organizational level. For instance, the consistent impact of value-driven motivations on team collaboration may inform approaches in team-based work environments, where shared goals and stable values are critical to long-term success. Organizations might benefit from fostering a culture that aligns with these underlying value orientations to enhance team dynamics and task effectiveness. The transient influence of attitudes towards specific technologies also suggests that simply promoting positivity towards a tool or process may not be enough to sustain engagement; instead, fostering an environment that resonates with employees’ fundamental values may yield more enduring, productive outcomes.

In addition, our results suggest that educational interventions aimed at motivating students in particular fields may need to go beyond short-term attitude adjustments. Programs designed to increase interest in specific fields, particularly for underrepresented groups, such as girls in STEM, often rely on brief exposures to role models or specific experiences intended to counteract cultural stereotypes. However, our findings may imply that lasting engagement may be more deeply rooted in personal values and personality traits rather than short-term attitudinal shifts. This insight raises questions about the efficacy of existing STEM motivation programs, which sometimes yield neutral or even negative effects on participants’ interest. It suggests that programs designed to stimulate interest in STEM fields might benefit from incorporating activities that align with participants’ intrinsic values, rather than focusing solely on fostering temporary positive attitudes toward technology.

7. Limitations and Future Research

This study has several limitations. First, it is a small-scale, exploratory study conducted within a single graduate-level course, which inherently limits the generalizability of the results. The participants are homogenous in terms of age, educational background, and academic major, all of which may restrict our ability to apply these findings to broader populations who may have varied levels of experience, attitudes, and motivations. Therefore, the findings may be most applicable to similar populations (e.g., young adults with a predisposition toward AI and societal issues). We understand the frustration that our readers may feel due to the small sample size. However, as we have addressed elsewhere, we believe the statistical tests are robust enough to provide a basis for further research. Future research should explore the role of self-transcendent goals and material aspirations in other collaborative, technology-enhanced educational contexts to test the robustness of our findings. Second, there is an unbalanced gender distribution among participants, with fewer male students represented in the study. This imbalance limits our understanding of how gender may interact with the attitudinal and personal variables under investigation. More diverse samples would allow researchers to examine whether gender moderates the relationships between personal motivators and collaborative, cognitive, and affective outcomes. Future research should strive for a more balanced gender representation to capture potentially significant differences and improve the applicability of findings across genders. Additionally, future studies could also employ mixed method approaches to investigate how gender dynamics influence group interactions. Qualitative methods, such as focus groups or observational studies, could provide deeper insights into how gendered communication styles and collaborative preferences affect learning outcomes in technology-enhanced educational settings.

Moreover, the study gathered self-reported, which may be biased, particularly on subjective constructs such as intrinsic motivation, cognitive flow, and collaboration. Future research could incorporate additional methods of data collection, such as observational data, behavioral analytics (e.g., task completion times or AI tool usage logs), or peer evaluations, could provide valuable triangulation and validation of self-reported measures.

Additionally, the study’s duration was relatively short, with data collected over a two-week design period. This limited timeframe may not fully capture the evolution of students’ attitudes, experiences, and motivations as they engage in more extended GenAI learning game design projects. Longer-term studies would allow for a deeper investigation into how initial attitudes and personal motivations fluctuate over time, especially as students encounter more complex challenges or achieve deeper familiarity with GenAI tools and collaborative dynamics. Therefore, future research should utilize longitudinal designs to investigate the persistence of the observed relationships over extended periods and across different stages of learning or project development. We also suggest future follow-up studies that incorporate qualitative methods, such as interviews or focus groups. These approaches could provide insights into students’ lived experiences, including specific challenges they encountered, their strategies for overcoming them, and how their perceptions of the technology evolved throughout the project.

We certainly would have liked to have conducted a study that strictly adheres to the canon of quantitative research. Although contentious, we would hold that it is better to do a small, imperfect study than not to do a study at all. (There are of course those who reasonably take the opposite position). Hence, we wish to be transparent about the severe limitations of both the size and diversity of the sample and ask the reader to bear this in mind when considering the study’s findings and generalizability.

8. Conclusions

This study explored the influence of personal value orientations, along with prior knowledge and attitudes toward technology and learning, on students’ positive experiences in designing AI-based learning games. By examining a variety of factors, the research highlights the significant roles self-transcendent goals and a desire for material benefits play in fostering positive collaborative, cognitive, and affective outcomes. Interestingly, while positive attitudes toward technology correlated with higher cognitive and affective engagement during the first week, their influence appeared to diminish over time.

Our findings provide a starting point for further exploration into how enduring values, rather than fleeting attitudes, shape students’ positive experiences with GenAI learning game design. These insights pave the way for future studies to investigate how value-oriented approaches can enrich collaborative and educational experiences with technology, potentially leading to the creation of more engaging and impactful AI-based learning tools. We specifically recommend taking action on turning transcendental and material goals into the design of GenAI. This can be carried out effectively by drawing on the expertise of educators who need to be included in the system design process. As such, teacher education can sensitize future educators to the value of this role as part of a well-balanced career.

If these findings are validated, this study could have meaningful implications for current educational initiatives aimed at changing students’ attitudes toward technology, particularly efforts to encourage more women to pursue STEM fields as mentioned above. Rather than emphasizing pro-technology instruction, focusing on value orientations might offer a more effective means of inspiring diverse participation in STEM programs. This perspective warrants further investigation, especially considering the significant investments currently being made in this area.