Generative AI is now part of how students learn. The Digital Education Council’s 2024 Global AI Student Survey of 3,839 university students in 16 countries found that 86% use AI in their studies and 54% use it at least weekly. Yet about half did not feel ready to use AI well, and 80% said their university’s integration of AI did not fully meet their expectations.
What AI changes
Tasks that only ask students to reproduce information, such as summaries, standard essays and recall questions, are the tasks AI tools complete most easily. Assessment built only on those tasks now tells a university less about what a student can actually do.
Why applied learning is the stronger response
Projects, live problems, presentations and peer discussion ask students to make judgements, explain their reasoning and respond to questions in real time. These methods also have a strong evidence base: active learning improves exam performance and reduces failure rates compared with traditional lecturing (Freeman et al., 2014). Employers want both kinds of capability: the World Economic Forum’s Future of Jobs Report 2025 lists AI and big data as the fastest-growing skills, alongside analytical and creative thinking.
Practical steps
- Teach AI use openly. Define where AI is allowed in each task and how students should acknowledge it.
- Assess the process, not only the product. Use milestones, drafts, reflections and short oral explanations.
- Use class time for application. Students can prepare with any tools they like, then apply and defend their thinking in class.
- Prepare students for productive struggle. In randomised classes, students learned more from active learning but felt they learned less, because the effort felt like confusion (Deslauriers et al., 2019). Explaining this helps students stay engaged.
A simple AI-use framework for every assignment
Students are rarely sure what is allowed, so each task should state its AI rule plainly. Many universities now use a small set of levels:
- No AI: for in-class tasks where the aim is to check what a student knows unaided.
- AI for ideas only: brainstorming and planning are allowed; the final work must be the student’s own.
- AI as a collaborator: AI may help draft or analyse, and students submit their prompts and explain what they changed.
- AI as the subject: students evaluate, test or critique AI output as part of the learning outcome.
UNESCO’s 2023 guidance on generative AI in education calls for this kind of human-centred, transparent approach, including building the AI competence of teachers as well as students (UNESCO, 2023).
Redesigning one assessment for an AI-rich world
Take a standard take-home essay. Instead of removing it, change what is assessed: students submit a short plan, a draft with tracked revisions, the final piece and a 300-word reflection on how they used sources and tools. A five-minute oral check, where the student explains one decision they made, confirms authorship far more reliably than detection software. The task now assesses judgement and reasoning, which is what employers value and what AI cannot supply on a student’s behalf.
Building faculty confidence
Students cannot be taught to use AI well by staff who have not had the chance to learn it themselves. Short, practical professional development that covers how the tools work, where they fail and how to redesign assessment is the fastest way to close the gap between student use and institutional policy.
Frequently asked questions
How many university students use AI in their studies?
The Digital Education Council’s 2024 Global AI Student Survey of 3,839 students in 16 countries found that 86% use AI in their studies and 54% use it at least weekly.
How should universities respond to students using generative AI?
By teaching responsible AI use openly, stating clearly where AI is allowed in each task, assessing the learning process as well as the final product, and using class time for application, discussion and explanation that show what students can genuinely do.
Can AI detection tools reliably identify AI-written work?
Detection tools can produce false positives and are easy to evade, so most guidance recommends redesigning assessment instead: drafts, reflections, in-class work and short oral explanations give stronger evidence of a student’s own understanding.
References
- Digital Education Council (2024). Global AI Student Survey 2024. Survey of 3,839 students in 16 countries. digitaleducationcouncil.com
- Freeman, S., Eddy, S. L., McDonough, M., Smith, M. K., Okoroafor, N., Jordt, H., & Wenderoth, M. P. (2014). Active learning increases student performance in science, engineering, and mathematics. Proceedings of the National Academy of Sciences, 111(23), 8410-8415. doi:10.1073/pnas.1319030111
- World Economic Forum (2025). The Future of Jobs Report 2025. Geneva: World Economic Forum. weforum.org
- Deslauriers, L., McCarty, L. S., Miller, K., Callaghan, K., & Kestin, G. (2019). Measuring actual learning versus feeling of learning in response to being actively engaged in the classroom. Proceedings of the National Academy of Sciences, 116(39), 19251-19257. doi:10.1073/pnas.1821936116
- UNESCO (2023). Guidance for Generative AI in Education and Research. Paris: UNESCO.