Page 66 - DP-2-2

Design+ Da Vinci AI Tutor in art history learning

not help them at all, while only 22% reported significant to match the content complexity with the user’s academic
benefits. This feedback highlights a disconnect between level. In addition, some students mentioned technical
the tutor’s capabilities and the students’ educational needs issues, such as slow loading times, which could disrupt the
or expectations. Furthermore, regarding future use of learning flow and affect user satisfaction. In fact, technical
tutors based on this experience, about 22% of respondents challenges were a recurrent theme, especially concerning
were very likely to use such a tool again, whereas another installation difficulties for Mac users. The comment, “Yes!
22% expressed reluctance or neutrality, suggesting varied I was never able to install it (Mac user),” was mentioned
satisfaction with the tutor. Students predominantly used multiple times, indicating a critical area for technical
text input (approximately 71%) to interact with the tutor, improvement. Ensuring that the tutor is accessible and
indicating a preference or greater ease with this method functional across various operating systems is essential for
compared to voice interaction or other modalities. This its successful integration into educational settings.
preference for text could reflect greater familiarity with Overall, qualitative feedback provided valuable insights
text-based interfaces or potential challenges in effectively into the strengths and limitations of the tutor, demonstrating
utilizing voice commands. its capability to foster enhanced comprehension, learner
The qualitative feedback from students regarding autonomy, and student engagement through personalized
their experiences also provides valuable insights into the interactions. However, key areas requiring targeted
performance of the system and areas for improvement. enhancements emerged, notably in system responsiveness,
Student responses to the survey questions about the most cross-platform accessibility, adaptive content complexity,
helpful aspects, least helpful or frustrating aspects, and and interaction simplicity. The systematic summary
any technical issues they encountered reveal a nuanced presented in Table 4 delineates these findings, highlighting
view of the current utility and limitations of the tutor. specific educational benefits, areas for technical and
For instance, students appreciated the ability to provide pedagogical improvement, and alignment with relevant
detailed and extended responses to their queries. One findings from prior studies, thereby facilitating an
student highlighted, “I liked how the AI elaborated with informed interpretation of the tutor’s efficacy and potential
every question I had, it went above and beyond to answer avenues for future development.
the original question.” This indicates that the AI’s depth
of information and ability to expand on topics were 5. Discussion
perceived as beneficial, particularly for those seeking The deployment and evaluation of the Da Vinci AI Tutor
comprehensive understanding and context. In addition, have demonstrated the potential of AI in higher education,
the personalization aspect, as reflected in another student’s particularly within the humanities. Through leveraging
comment, “I liked Da Vinci’s personality,” suggests that generative models and a historically informed avatar,
the AI’s character design added a unique and engaging this initiative sought to enhance student engagement,
dimension to the learning experience. personalize learning, and improve accessibility. However,
the study also highlighted critical limitations and areas
Feedback on areas for improvement was varied, indicating
diverse expectations and experiences among the users. One for refinement. The following subsections outline
the theoretical implications, managerial or policy
respondent noted, “My inquiries were through ChatGPT recommendations, and directions for future research,
and my answers were in response to that program. Really drawing on the lessons learned from this implementation.
did not have an opportunity to work with Art AI program Regardless, the study contributes significantly to
through the university. Will attempt after this semester is the emerging field of AI applications in humanities
over.” This comment points to potential confusion or lack education, specifically addressing the need for adaptive,
of clarity regarding how to engage with the tutor effectively. personalized learning experiences. The tutor demonstrates
Another student, unable to install the program due to being the feasibility of leveraging GAI, XR environments, and
a Mac user, highlighted a significant accessibility barrier. adaptive learning models to effectively enhance student
This issue underscores the need for broader compatibility engagement, comprehension, and critical thinking skills
and easier installation processes to ensure all students can within art history coursework. By integrating cognitive
access the tool regardless of their hardware.
scaffolding strategies, responsive dialogue, and immersive
Responses to what were least helpful or frustrating virtual contexts, the tutor aligns with established
about the tutor included comments about the overload of educational theories, including Vygotsky’s zone of
information, with one student saying it provided “Way too proximal development and Kolb’s experiential learning
much information.” Others felt the responses were “Too theory, effectively supporting student learning beyond
basic for my needs,” suggesting a disparity in the ability mere content delivery.

Volume 2 Issue 2 (2025) 16 doi: 10.36922/dp.8365