The numbers Students aren't waiting for permission. HEPI's survey of UK undergraduates found 95% now use AI in at least one way.

Understanding hasn't kept pace. The Digital Education Council's latest research puts AI literacy at an early stage for students and staff alike. And there are early signs of dependence: 21% of respondents said they find it harder to work through problems without AI, and 20% said they rely on it to produce better work.

Staff are no more confident. In a report from the American Association of Colleges and Universities and Elon University, 62% of faculty said graduates weren't prepared to use generative AI well. 68% said their own institution hadn't prepared them to teach with it.

So adoption has run well ahead of understanding. That's the core argument in a recent University World News commentary from Petra Recter of Oxford University Press, and the figures back it up.

Why students reach for it Time, mostly. Nearly nine in 10 incoming UK students expect to work during term, and under a quarter think they'll have enough money to cover living costs. In the US, 67% of undergraduates work for pay and around one in five are parents.

If you're juggling shifts, childcare and a seminar deadline, a tool that saves you an hour is hard to turn down. I'd do the same.

And the use is broader than the headlines suggest. Students use AI to research unfamiliar topics, generate ideas, organise their thinking, refine drafts and get feedback. Essay- writing is only one slice of it.

What OUP is offering OUP's answer is a six-part AI literacy module, built from interactive video and assessments. Lecturers can assign it to students through Oxford Insight, Oxford Learning Link and Oxford Trove. It covers:

  • what generative AI is, with common tools and basic ethics
  • AI and academic integrity, including how to document use and follow institutional policy
  • risks such as hallucinations and misinformation, set against benefits like speed and creativity
  • critical thinking techniques for checking outputs for errors and bias
  • prompting, including refining prompts through iteration
  • a course-specific AI activity written by a subject expert

The foundation module was authored by Dr Anand Rao of the University of Mary Washington and Stefan Bauschard of Union College. OUP has also run faculty webinars on AI use.

My take Let's be upfront about what this article is. It's a commentary by OUP's global director of content and digital strategy, and it promotes an OUP product. That doesn't make it wrong. The outline is sensible, and the course-specific activity is the strongest part. Literacy sticks when students apply it to their own subject.

But I'd push back in a few places.

A module can be ticked off. Those dependence figures, where one in five students struggle to think through a problem without AI, won't shift because of a video and a quiz. They shift when assessment design asks students to show their thinking, and when tutors model good use week in, week out.

Staff come first. If 68% of faculty feel their institution hasn't prepared them, a student module only papers over the gap. Students take their cues from the people teaching them. Train the staff and the student literacy follows.

Prompting is the easy bit. It gets a whole section here. The harder skill is knowing when to leave AI out entirely. I've used the escalator and stairs comparison before: the escalator gets you to the top quicker, but your legs don't get any stronger.

Who teaches students to question the tools? Literacy built by a publisher, delivered inside that publisher's AI-enabled platforms, will naturally lean towards using AI well. Fair enough. Students also need room to ask whether a tool should be used at all, and who benefits when it is. That's the same question the Lawrence school district's AI committee said it never got to ask.

Where this leaves us Recter is right that AI literacy can't wait. Employers increasingly want AI skills, yet 51% of new graduates say they aren't actively building them.

The same applies well beyond universities. The challenge extends to colleges, apprenticeship providers and schools, where AI literacy also needs attention. An off-the-shelf module can start the conversation. The lasting work happens in assessment design and in giving our own staff the confidence to teach it.

So who's teaching AI literacy where you work? A bought-in module, your own staff, or nobody yet?

Source How do institutions embed AI literacy in higher education?, Petra Recter, University World News, 17 October 2026.