INTERNATIONAL JOURNAL OF ARTIFICIAL INTELLIGENCE
ISSN: 2692-5206, Impact Factor: 12,23
American Academic publishers, volume 05, issue 07,2025
Journal:
https://www.academicpublishers.org/journals/index.php/ijai
page 103
WHEN AI ISN’T ENOUGH: THE UI/UX EXCEPTION IN AI-SUPPORTED
LEARNING AT JAPAN DIGITAL UNIVERSITY
Boboyev Lochinbek Boymurotovich
PhD, head of IT department of Japan Digital University,
,
Ziyodullayev Amirbek Akmalovich
student of Japan Digital University
ziyodullayevamirbek238@gmail.com
Abstract.
Artificial Intelligence (AI) has revolutionized educational methodologies by
providing adaptive learning environments that enhance student performance in technical
disciplines. At Japan Digital University (JDU), a phased integration of AI tools was
implemented across the IT curriculum, including courses in programming, data management,
and project planning. The results demonstrated consistent academic improvement across AI-
supported subjects. However, the User Interface and User Experience (UI/UX) Design course
remained an exception, being taught without AI involvement. This study analyzes academic
outcomes, student engagement, and cognitive-emotional responses in both AI-integrated and
traditional learning contexts. Findings suggest that while AI optimizes structured learning, it
falls short in creative, subjective domains like UI/UX, where empathy, aesthetic judgment, and
human-centered thinking are paramount. The research concludes that educational institutions
must adopt a balanced, discipline-sensitive approach to AI integration to preserve essential
human elements in learning.
Keywords:
Artificial Intelligence, UI/UX Design, Creative Education, Adaptive Learning,
Student Performance, Japan Digital University, Human-Centered Learning, Educational
Technology, Cognitive Engagement, Pedagogical Strategy
I. Introduction.
Artificial Intelligence (AI) is no longer a futuristic concept but a transformative force in modern
education. Over the past decade, AI has become deeply embedded in teaching and learning
processes worldwide, bringing significant shifts in how students engage with content, how
instructors manage classrooms, and how academic performance is measured. Through
intelligent tutoring systems, adaptive feedback platforms, and AI-based learning analytics,
institutions now offer students a more personalized, efficient, and scalable educational
experience.
According to Gligorea et al. (2023), adaptive AI platforms significantly improve retention and
student engagement in online and blended education settings by allowing learners to move at
their own pace and receive timely, individualized support [1]. AI tools not only provide
automated feedback and code corrections but also help students detect errors, understand
theoretical concepts, and manage learning tasks independently. Yet, with these advancements
come new concerns. Scholars such as Zhai et al. (2024) argue that excessive dependence on AI
systems may inhibit students' long-term cognitive development by replacing effortful thinking
INTERNATIONAL JOURNAL OF ARTIFICIAL INTELLIGENCE
ISSN: 2692-5206, Impact Factor: 12,23
American Academic publishers, volume 05, issue 07,2025
Journal:
https://www.academicpublishers.org/journals/index.php/ijai
page 104
with machine-generated shortcuts [2].
As demonstrated in previous research conducted at Japan Digital University (JDU), the
integration of AI-supported learning into the IT curriculum followed a clearly defined, phased
structure and was closely observed through the academic progress of students admitted in 2021.
The study covered three consecutive semesters, each representing a distinct stage in the AI
integration process:
Semester 1 – Traditional instruction without the use of AI tools;
Semester 2 – Partial integration of AI, including video lectures, code assistants, and basic
chatbot support;
Semester 3 – Full-scale implementation of AI tools across most technical subjects. [6].
The results revealed a steady improvement in student performance. In subjects such as PHP &
SQL, Python, Object-Oriented Programming, and Project Management, academic outcomes
improved significantly. The use of AI tools—including intelligent suggestions, automated
feedback, and adaptive instructional systems—enabled students to grasp complex topics more
quickly and approach practical tasks with greater confidence [6].
However, amidst this digital transformation, one course remained untouched by AI: User
Interface and User Experience (UI/UX) Design. While technical subjects saw increasing
automation, UI/UX was consistently delivered using traditional methods—manual prototyping,
live critiques, instructor-led design reviews, and peer collaboration. This pedagogical decision
was intentional. UI/UX, unlike programming, is rooted in human emotion, intuition, and social
sensitivity. It requires the designer to think not like a machine, but like a user—to empathize, to
anticipate needs, and to craft experiences that resonate across cultures and contexts.
As Saini & Ahmed (2023) point out, although AI can assist in generating basic layouts or
optimizing visual contrast, it lacks the emotional intelligence and cultural awareness needed to
design interfaces that truly connect with users [3]. The same point is emphasized by UX experts
like Malik (2023), who argues that AI still fails to answer key questions like "Does this design
feel intuitive?" or "Will it frustrate the user?"—questions that rely on context, empathy, and
creativity rather than data-driven logic [4].
Furthermore, Lu et al. (2024) stress that collaborative creativity, emotional nuance, and
aesthetic judgment remain outside the scope of current AI systems, especially in visual and
interaction design [5]. When comparing student outcomes, a clear trend emerges: while
performance steadily improved in AI-enhanced courses, the UI/UX course maintained a higher
proportion of failing or underperforming students. This suggests that students, accustomed to
automated support in other areas, may have struggled when required to think independently,
make subjective decisions, and defend their creative choices without digital assistance.
II. Main part.
INTERNATIONAL JOURNAL OF ARTIFICIAL INTELLIGENCE
ISSN: 2692-5206, Impact Factor: 12,23
American Academic publishers, volume 05, issue 07,2025
Journal:
https://www.academicpublishers.org/journals/index.php/ijai
page 105
2.1. Academic Trajectory Across AI Integration Phases
To understand the impact of AI on student outcomes, Japan Digital University implemented a
phased integration of AI into its IT curriculum. Each of the three semesters represented a
distinct level of AI adoption and offered a window into how students responded to new
instructional technologies [6].
In Semester 1, teaching relied solely on traditional lectures without any AI support. Students
had limited access to supplementary resources, and classes were large, making individual
assistance difficult. As a result, many students struggled in subjects such as PHP & SQL and
Object-Oriented Design (OOD). The lack of personalized support contributed to a high rate of
academic failure [6].
Diagram 1. Grade distribution during the 1st semester, reflecting the traditional, non-AI-
supported learning environment. A significant number of students (77) received failing grades
(F), highlighting the limitations of conventional teaching methods in large IT classes.[6]
The second semester marked the transition phase. AI-based tools such as video lectures,
basic coding assistants, and chatbot interfaces were introduced. Instructors uploaded pre-
recorded lectures to the online platform, and classroom time was repurposed for Q&A sessions
and collaborative problem-solving. However, both students and instructors faced adjustment
difficulties: AI platforms were underutilized due to inexperience, and many students struggled
to formulate structured questions for the AI tools. Nonetheless, the environment became more
flexible, and some students began to benefit from on-demand access to learning materials and
automated feedback.[6]
INTERNATIONAL JOURNAL OF ARTIFICIAL INTELLIGENCE
ISSN: 2692-5206, Impact Factor: 12,23
American Academic publishers, volume 05, issue 07,2025
Journal:
https://www.academicpublishers.org/journals/index.php/ijai
page 106
Diagram 2. Grade distribution in the 2nd semester following the partial integration of AI tools
and video lectures. While the number of F grades remained high (84), the increased number of
A and B grades indicates early signs of improvement due to AI-assisted learning.[6]
In the third semester, AI integration reached its full potential. Nearly all technical
subjects—Project Management, Python, Django, .NET, and OOP—were supported by
intelligent systems that could assess student work, offer real-time corrections, and provide
adaptive feedback. Students showed increased confidence, problem-solving efficiency, and task
independence. Teachers shifted from direct instruction to facilitation roles, supporting higher-
level
discussions
and
personalized
mentoring.[6]
Diagram 3.
Grade distribution during the 3rd semester, reflecting the full adoption of
AI tools in teaching. A marked improvement is observed, with 50 students earning A grades
and a reduced failure count (42), showcasing the effectiveness of AI-supported pedagogy.[6]
INTERNATIONAL JOURNAL OF ARTIFICIAL INTELLIGENCE
ISSN: 2692-5206, Impact Factor: 12,23
American Academic publishers, volume 05, issue 07,2025
Journal:
https://www.academicpublishers.org/journals/index.php/ijai
page 107
Table.
Comparative summary of student performance across the 1st, 2nd, and 3rd
semesters. The data shows a steady improvement in academic outcomes as AI tools became
more integrated into the IT curriculum.[6]
2.2. UI/UX Design as a Non-AI Course
Unlike other courses in the IT curriculum, the UI/UX Design course was
deliberately excluded from AI enhancement. This subject was taught through traditional
methods: instructor-led seminars, sketch-based prototyping, and peer feedback sessions.
Students were required to develop user personas, create design mockups manually, and
justify their decisions in verbal presentations. No AI platforms were used to suggest
layouts, generate wireframes, or interpret user feedback. [6].
This approach was based on pedagogical reasoning. UI/UX education emphasizes
human empathy, visual perception, storytelling, and emotional resonance—dimensions
that are difficult, if not impossible, to simulate with current AI systems [4]. Creativity in
UI/UX design often involves ambiguity, emotional logic, and social context that do not
lend
themselves
well
to
data-driven
processing
INTERNATIONAL JOURNAL OF ARTIFICIAL INTELLIGENCE
ISSN: 2692-5206, Impact Factor: 12,23
American Academic publishers, volume 05, issue 07,2025
Journal:
https://www.academicpublishers.org/journals/index.php/ijai
page 108
[5].
Diagram 4
presents the grade distribution in the UI/UX course, which remained AI-free.
The results show a concentration of mid-range grades, with fewer students achieving top marks.
Moreover, students faced unique challenges. Those accustomed to the instant support and
feedback from AI platforms in other courses felt uncertain in UI/UX tasks that demanded
subjective judgment. They often hesitated to finalize design decisions or struggled to articulate
their choices during critiques. The absence of automated guidance required a deeper level of
personal responsibility and introspection.
2.3. Patterns and Disparities in Student Performance
While AI-assisted courses showed steady progress over three semesters, UI/UX
remained an outlier both in structure and results. Interestingly, although AI-supported courses
like Project Management and Python exhibited higher A and B rates, the number of F grades
was still notable. This reveals that AI, while helpful, is not a universal solution; student
motivation, participation, and foundational knowledge remain critical to academic success
[1][2][6].
In contrast, the UI/UX course generated more C and D grades, suggesting that students
struggled to meet design expectations without structured digital help. However, the number of
outright failures was not significantly higher than in technical subjects. This implies that the
absence of AI did not necessarily lead to greater academic failure, but it did expose a deeper
gap in students’ creative independence, visual reasoning, and critical reflection [5][6].
INTERNATIONAL JOURNAL OF ARTIFICIAL INTELLIGENCE
ISSN: 2692-5206, Impact Factor: 12,23
American Academic publishers, volume 05, issue 07,2025
Journal:
https://www.academicpublishers.org/journals/index.php/ijai
page 109
Unlike technical subjects where errors can be clearly diagnosed and corrected with AI
assistance, UI/UX assessments rely on subjective interpretation, user empathy, and design logic.
Many students, accustomed to binary correctness and automated validation, struggled when
faced with open-ended tasks and ambiguous outcomes. Without AI-generated scaffolding, some
students hesitated to take creative risks or to engage fully in peer critiques, fearing judgment
without algorithmic reinforcement [4][6].
This disparity highlights the different cognitive and emotional demands across
disciplines. AI is highly effective in domains that require accuracy, structure, and repetition.
But in creative courses like UI/UX, students must learn to rely on human feedback, intuition,
and personal judgment—competencies that cannot be automated [5].
2.4. Student Perception and Engagement
While AI tools brought significant academic improvements in technical subjects, their
absence in the UI/UX course revealed deeper differences in how students perceived their
learning experience. In AI-supported courses such as Python, SQL, or Project Management,
students reported a higher sense of control over their progress. The availability of instant
feedback, automated assistance, and structured learning materials allowed them to approach
assignments with greater confidence and independence [1][3]. These tools also reduced anxiety
by offering quick validation and troubleshooting during challenging tasks [3].
In contrast, the UI/UX course, which lacked any AI integration, demanded a different
cognitive and emotional approach. Students were expected to justify their design choices during
live critiques, defend their decisions based on subjective feedback, and manage open-ended
tasks without concrete "right" answers. Many found this transition difficult. Having become
accustomed to the clarity and support of AI in previous courses, they struggled to adapt to an
environment where creativity, intuition, and human judgment were central [4][5].
Although no formal survey was conducted, informal interviews and classroom
observations indicated that students were less confident when engaging with design challenges
in UI/UX. Some expressed hesitation in finalizing decisions, while others avoided experimental
or unconventional ideas due to fear of subjective criticism [6]. This emotional burden impacted
their ability to express creativity and take risks—elements that are essential to the learning
process in design education [5].
Overall, the contrast between the AI-supported and non-AI-supported environments
highlighted how digital assistance not only influences academic performance, but also shapes
students’ emotional engagement, risk tolerance, and confidence in independent thinking. The
case of UI/UX confirms that in creative disciplines, the absence of AI can expose important
gaps in students’ readiness for open-ended, human-centered learning [4][5].
2.5. The Limits of AI in Creative Education
Despite its advantages in optimizing and personalizing technical education, AI has clear
boundaries in creative disciplines. It can replicate patterns, generate layouts, and automate
design suggestions, but it cannot comprehend emotional nuance, cultural symbolism, or user-
centered narratives [4][5]. In UI/UX, students must anticipate user intent, express personality
INTERNATIONAL JOURNAL OF ARTIFICIAL INTELLIGENCE
ISSN: 2692-5206, Impact Factor: 12,23
American Academic publishers, volume 05, issue 07,2025
Journal:
https://www.academicpublishers.org/journals/index.php/ijai
page 110
through design, and align visual elements with psychological effects—tasks that require human
judgment [5].
Instructors in design education must nurture qualities that AI cannot simulate: empathy,
ethical reasoning, storytelling, and aesthetic sensitivity [4][5]. While AI may serve as a
supportive tool in basic design stages (e.g., checking contrast or accessibility), it cannot replace
the mentorship, critique, and emotional development central to creative learning [5].
Table2.
Comparison of tasks AI can support versus tasks that require human creativity,
intuition, and judgment in UI/UX education.
III. Conclusion.
The integration of artificial intelligence into higher education has opened new pathways
for adaptive, student-centered learning—especially in technical disciplines. As demonstrated
through the phased implementation at Japan Digital University, AI tools significantly enhanced
academic outcomes in programming-related courses such as PHP & SQL, Python, and Project
Management. Students benefited from immediate feedback, round-the-clock access to guidance,
and individualized pacing. The progression from traditional instruction in Semester 1 to full AI
integration in Semester 3 resulted in a clear improvement in academic performance, reduced
failure rates, and increased learner autonomy.
However, the case of the UI/UX Design course illustrates that AI is not universally
applicable across all areas of education. This subject remained outside the scope of AI support,
and its pedagogical structure relied entirely on human interaction, subjective judgment, and
creative inquiry. While students still achieved reasonable academic results, their confidence and
satisfaction levels were often lower. The absence of automated assistance exposed difficulties
in emotional reasoning, decision-making, and visual communication—skills that current AI
systems cannot replicate.
These findings suggest that while AI is highly effective in optimizing structured, logic-
driven tasks, its capabilities are limited when it comes to fostering creativity, empathy, and
ethical thinking. In disciplines like UI/UX, education must remain human-led, with a focus on
mentorship, exploration, and emotional development. Future applications of AI in education
should therefore be discipline-sensitive: used where appropriate and withheld where it might
hinder critical human capacities.
INTERNATIONAL JOURNAL OF ARTIFICIAL INTELLIGENCE
ISSN: 2692-5206, Impact Factor: 12,23
American Academic publishers, volume 05, issue 07,2025
Journal:
https://www.academicpublishers.org/journals/index.php/ijai
page 111
Ultimately, a balanced approach is needed—one that combines the efficiency of AI with
the irreplaceable strengths of human educators. As educational institutions continue to embrace
digital transformation, thoughtful integration strategies will be essential to ensuring that all
learners benefit, regardless of subject matter.
REFERENCES:
1. Roll, I., & Wylie, R. (2016). Evolution and revolution in artificial intelligence in education.
International Journal of Artificial Intelligence in Education, 26(2), 582–599.
https://link.springer.com/article/10.1007/s40593-016-0110-3
2. Zhang, J., & Zheng, Y. (2020). Challenges and opportunities of AI for creative thinking in
education. Educational Technology Research and Development, 68(2), 857–860.
https://link.springer.com/article/10.1007/s11423-020-09779-0
3. Nguyen, H. N., Dang, T., & Nguyen, V. T. (2020). Interface Design for HCI Classroom:
From Learners' Perspective. arXiv. https://arxiv.org/abs/2010.01651
4. Chen, W., Yu, H., & Chang, C. (2021). Adaptive learning platforms: Impacts on retention
and
performance
in
STEM.
Frontiers
in
Education,
6,
702178.
https://www.frontiersin.org/articles/10.3389/feduc.2021.702178/full
5. Anantrasirichai, N., & Bull, D. (2020). Artificial intelligence in the creative industries: A
review. arXiv. https://arxiv.org/abs/2007.12391
6. Ziyadullaev, A. (2025). The application of artificial intelligence in education: A case study
at
Japan
Digital
University
https://global-science-
review.com/ojs/index.php/gsr/article/view/1928
