Smart Glasses in the Exam Hall: Ofqual's Cheating Panic Is Actually a Teaching Opportunity

England's exam regulator is worried about AI specs and hidden earpieces turning GCSEs into open-book tests. Scottish schools are watching closely. But buried inside the panic is a more useful question: what are we actually trying to assess, and does AI change that answer?

Ofqual, the exam watchdog for England and Wales, has flagged smart glasses, hidden earpieces, and AI tools as a growing threat to the integrity of GCSE and A-level exams. The concern is straightforward: a student wearing the right pair of specs could, in theory, stream a question to an AI, receive a whispered answer, and write it down — all without the invigilator noticing a thing. It is not science fiction. The hardware exists today, off the shelf, at consumer prices.

Scotland runs its own qualification system through the Scottish Qualifications Authority, but the pressure is identical. SQA Nationals, Highers, and Advanced Highers all rely on the same fundamental assumption that underpins every exam since Aristotle: the person in the seat is doing the thinking. That assumption is now genuinely fragile, and no Scottish school, college, or university has fully solved it yet. Education Scotland's ongoing curriculum review has touched on digital skills and AI literacy, but a coherent national policy on AI in assessment remains a work in progress.

The kneejerk response — ban everything, strip pupils of devices, treat every candidate as a suspect — is already showing its limits in England. According to research published by JISC, the UK's digital education charity, blanket technology bans in exam settings are increasingly difficult to enforce as wearable tech becomes smaller and harder to detect. Meanwhile, teachers, pupils, and parents are asking a more fundamental question: if an AI can pass a Higher Chemistry paper, what exactly is that paper testing?

That question is uncomfortable, but it is the right one. The most forward-thinking educators in Scotland are not asking how to keep AI out of the room — they are redesigning what they ask pupils to do. Project-based assessment, oral examinations, and tasks that require demonstrating reasoning in real time are all harder to outsource to a language model. The University of Edinburgh's own assessment working group has been examining AI-resilient evaluation methods since 2023, and several Scottish secondary schools have piloted portfolio-based approaches that value process over product. These are not soft alternatives; they are harder to fake and arguably test more of what employers and universities actually want.

For Scottish headteachers and education leads, the Ofqual warning is a useful prompt rather than a crisis. The practical task right now is threefold: update your acceptable-use policy to specifically name AI tools and wearable technology; start a staff conversation about which assessments in your school are genuinely AI-resilient and which are not; and engage with whatever guidance SQA publishes this academic year. The Scottish Government's AI in Education working group is expected to issue updated guidance before the 2025–26 session. Getting ahead of it, rather than scrambling to catch up, is the difference between a policy and a panic.

What this means for you: If you run a tutoring company, an independent school, or any education-adjacent business in Scotland, AI assessment integrity is the conversation your clients are already having — and they will pay for clarity. Consider whether your own service policies address AI use explicitly; vague terms now create disputes later. If you employ or train young people, it is also worth knowing that the skills AI cannot replicate — critical reasoning, verbal explanation, adaptive problem-solving — are exactly what you should be testing at interview.