AI Tools for Student Learning : Exploring the Impact of KAI on Learning and Its Potential to Enhance Learning for University Students

Sammanfattning

The integration of Artificial Intelligence (AI) into higher education is reshaping how students interact with learning platforms. This study explores the pedagogical and technical impact of KAI (Kollektiv AI), the AI assistant in the Swedish platform Kollin.io, focusing on university students’ self-study and academic engagement. Using surveys and interviews with KTH students, the research examines how AI can support learning and what limitations emerge. Findings show that KAI offers helpful, course-aligned feedback, especially during exam preparation. However, students noted issues with vague answers, multi-part questions, and sensitivity to phrasing, prompting some to prefer broader tools like ChatGPT. These challenges reflect KAI’s underlying architecture, which uses commercial language models via Vercel’s AI SDK and relies on precise input. While technically promising, its effectiveness hinges on design choices that balance usability, accuracy, and pedagogy. The study highlights how AI assistants like KAI can complement traditional learning, if developed with user-centered design and continuous refinement in mind.