Abstract:

Video-based learning has become a core element of university education, yet students often struggle to identify which videos best support their needs at different stages of the learning process. Although large collections of instructional videos are available, students frequently report uncertainty about which resources to choose, how to evaluate their relevance, and how to integrate them into their study routines. To better understand these challenges and to explore the potential role of artificial intelligence in addressing them, this study places a targeted student survey at the centre of its analysis.

The survey was designed to capture four dimensions:
(1) students’ current use of videos,
(2) the difficulties they encounter when searching for suitable content,
(3) their perceptions of potential AI-based recommendations, and
(4) the expected impact of such tools on their learning experience.

Results show that videos are used frequently and are viewed as highly valuable for clarifying concepts, reviewing before assessments, and compensating for gaps in understanding. However, many students report that finding the “right” video is often difficult. They describe challenges such as excessive options, unclear titles, mismatches between video complexity and their own level, and uncertainty about where to start when approaching a new topic.

Importantly, the survey reveals strong student interest in AI-generated recommendations. Learners indicate that personalised suggestions could help them locate short explanations for quick revision, tutorials aligned with exercises, clearer explanations of difficult topics, or real-world examples that reinforce theoretical content. Students also believe that recommendations tailored to their progress would save time, reduce frustration, and support more autonomous learning habits.

Building on these findings, the study argues that AI-driven video recommendation systems have the potential to address the specific issues identified by students. By analysing video content and aligning it with learners’ preferences and difficulties, such systems could streamline access to relevant materials and enhance the overall learning process. The survey thus provides concrete empirical evidence that can guide the design of future educational tools and inform pilot implementations in university courses.

Overall, the results highlight both a clear pedagogical need and strong student readiness for AI-supported video navigation. The study contributes a student-centred perspective to the discussion on AI in education and outlines practical avenues for developing and evaluating recommendation systems that improve the accessibility, efficiency, and personalisation of video-based learning.

Keywords:

Video-based learning, higher education, artificial intelligence, recommendation systems.