Abstract:

Providing personalised feedback in higher education is a persistent challenge, particularly in courses with large student groups, where instructors must guide each learner’s process in an individualised way. Developing students’ capacity for self-assessment and reflection is therefore essential, as these skills foster autonomy, creativity, and the ability to iteratively improve their work. While recent advances in artificial intelligence (AI) have demonstrated significant potential for supporting feedback on written tasks, there has been limited exploration of AI systems that assist students in evaluating visual and creative outputs, such as design projects.

This study presents the development and evaluation of an AI-based tool designed to provide formative feedback on students’ design work through image analysis. The project is a joint initiative between the Universitat Politècnica de Catalunya and the Valencian Research Institute for Artificial Intelligence (VRAIN) from the Universitat Politècnica de València. The system analyses students’ submitted images and, following evidence-based principles of effective feedback, generates constructive comments that help learners critically reflect on their design decisions and identify possible improvements in their work.

The tool was implemented in a classroom setting with 58 undergraduate students enrolled in a Graphic Design and Communication course. After interacting with the system, participants completed a structured survey using five-point Likert scales, and their conversations with the AI were analysed to identify feedback patterns. Quantitative results indicate that students evaluated the tool positively, considering the feedback useful (M = 4.00). Many noted that the interaction helped them improve their design knowledge (63%), reported that communication was easy (60%), and expressed interest in using similar tools in future projects (63%). Nevertheless, over half of respondents (55%) expressed disapproval of teachers using AI for assessment purposes, favouring its use as a formative support tool instead. Interaction analysis revealed that students engaged actively with the system, making frequent feedback requests, although their prompts were often brief and tended to be concise.

Overall, the findings highlight the potential of AI as a learning companion that fosters reflective practice, self-regulation, and engagement, especially in contexts where personalised tutor feedback is constrained by large class sizes. The study contributes to a deeper understanding of how students engage in self-reflection through AI-based feedback systems in creative learning contexts, informing future research that seeks to integrate AI to support formative learning.

Keywords:

Artificial Intelligence, Educational Technology, Formative Assessment, Higher Education, Reflective Learning.