Three years ago, I walked into a seminar room and announced a new rule: no generative AI, ever. I explained, with the certainty of someone newly arrived from the corporate world, that AI would rot their critical thinking. A student in the back row raised a hand.
“Professor, I used ChatGPT last night to rewrite my CV for an internship application. Should I delete that too?”
The room went quiet. I had no answer. I was from Mars, issuing edicts. She was from Venus, already using the tool to navigate the real world.
That moment was my first clue that our AI conversation was broken. But it took me another year to understand how broken.
Students are not sneaking AI into essays because they are lazy. They use it because their lives are full of tasks academia ignores: drafting emails to employers, decoding accommodation contracts, preparing for interviews they never learned to ace. More than half (53 per cent) of students use generative AI for non‑academic purposes – from writing personal statements to planning budgets – while fewer than one in five feel their university has given any ethical guidance, according to research. This is not cheating. It is self‑directed professional development in a vacuum of institutional support.
- Dear educators, Gen Z here. Could you please teach us like it’s 2026?
- An insider’s guide to how students use GenAI tools
- Assessing critical thinking in critical times
Meanwhile, faculty remain anchored on Mars. We worry about plagiarism detectors and policy loopholes. Of UK and European academics, 41 per cent have never used generative AI in any form, and 29 per cent have actively blocked access to tools in their virtual learning environments, a 2025 survey showed.
When I asked colleagues why, the answers were honest: “I don’t know how it works.” “I’m afraid of looking stupid.” We have turned a competence gap into moral panic – and our students are paying the price.
And it’s measurably damaging. A study of teacher‑student relationships found that 62 per cent of students felt their lecturers viewed them with automatic suspicion, and 57 per cent reported hiding their AI use even when they believed it was legitimate. This highlights a failure of mutual transparency: students declare AI use but faculty rarely do the same. The result is a classroom of silent distrust – the opposite of the open culture we claim to foster.
But here is the discovery that should pull us back from the edge. Faculty and students share the same core concern, according to a 2025 comparative study: that AI might erode authentic learning. Notably, students were more worried than faculty about becoming over‑reliant on AI. The people we imagine are running wild with chatbots are actually more cautious than we are. The gap is not in values. It is in visibility and trust.
What would it take to close that gap? Not more surveillance. Not longer policy documents. Three shifts, starting Monday:
Shift one: admit we are beginners
Every university now requires digital literacy training for students. Faculty need the same – without shame. A half‑day workshop on prompt crafting, bias spotting and assignment redesign should be mandatory. At my university, we started a monthly “AI lunch” where faculty bring their failures. It has done more to change practice than any policy.
Shift two: write classroom guidelines together
Last term, I gave my students a blank Google Doc and asked them to draft three rules for AI use. They produced: “Cite any AI‑generated text”, “No copy‑paste without rewriting” and “Ask the professor if unsure.” Then they enforced those rules on each other. Compliance became peer pressure, not policing.
Shift three: design assessments that reward process, not polish
If an assignment can be completed by AI in 10 seconds, it is a bad assignment. Replace final essays with oral defences, iterative drafts or live simulations where students must explain their choices under pressure. One colleague now asks students to submit their ChatGPT conversation history alongside their final paper – not to catch them, but to see how they refined the output. That is pedagogy, not prosecution.
The real question is not how to stop students from using AI. They will use it – on Venus, in their careers, long after they have forgotten our threats. The real question is whether we want to be Martians shouting into space or fellow travellers who learn to navigate the same galaxy.
Not obedience. Curiosity.
Dina Kamel is a teaching fellow at the University of Portsmouth. An LLM was used for English language review.
If you would like advice and insight from academics and university staff delivered direct to your inbox each week, sign up for the Campus newsletter.
comment