Student perceptions on the use of AI for learning and assessments: a "love-hate relationship"

Artificial intelligence (AI) is reshaping higher education, yet adoption and AI literacy remain uneven across institutions and student groups. Research in the UK indicates that variation in prior educational experience, institutional support, and confidence with emerging technologies shapes student AI literacy (O’Dea et al., 2024). Sector data show rapid growth in generative AI use: student use of tools such as ChatGPT for assessment rose from 53% to 88% between 2024 and 2025, with common applications including explaining concepts, summarising readings, and suggesting research ideas (Freeman, 2025). National reports also document routine uses for note-making, revision, drafting, and interpreting feedback, alongside concerns about accuracy, access, and unclear guidance (Attewell, 2025). In addition, widening participation students report lower digital confidence and less consistent use of formal digital resources, which may limit their ability to benefit from AI-supported learning (Zhou, Fang & Rajaram, 2025). This poster presents findings from a recent study at LSBU that explores perceptions and usage of generative AI tools for learning and assessment among students who typically come from non-traditional backgrounds and with varied digital capabilities. The study combined surveys and focus groups to capture the student experience. Preliminary results indicate that LSBU students have become more familiar with AI and are adopting more strategic and critical approaches, using AI for idea generation, feedback seeking, and understanding complex concepts rather than accepting outputs without evaluation. Despite this shift, disbelief and uncertainty persist concerning the capabilities and appropriate use of AI for learning and assessment. The study findings mirror sector reports of increased generative AI use while also reflecting shared concerns about accuracy, access, and guidance (Freeman, 2025; Attewell, 2025). In conclusion, the study highlights the need for consistent institutional guidance, clear ethical frameworks, and structured training to support equitable, pedagogically sound use of generative AI in higher education.