Ofqual warns of far tougher scrutiny on AI in coursework

TL;DR:

• Ofqual chief Ian Bauckham says written coursework will face “far, far more scrutiny” under reformed qualifications to prevent the “AI-fuelled subversion of assessment”.
• The warning comes ahead of the first three V Level subjects launching next year.
• He cautioned that AI marking is “not actually as good as you might think”, citing hallucinations and bias.

England’s assessment regulator is drawing a line on AI in the exam system. Ofqual chief regulator Ian Bauckham used the Festival of Education to warn that any proposals for written coursework in new and reformed qualifications will face “far, far more scrutiny”, arguing schools and colleges cannot “normalise the idea that AI-generated output is a substitute for genuine human endeavour”. It is one of the most concrete UK policy responses yet to generative AI in education.

Assessment design under pressure

Bauckham pointed to universities already reworking assessment to cope — citing UCL Law School’s shift towards in-person formats to make exams “AI-proof” — and framed the coming V Levels, the vocational qualifications for 16- to 18-year-olds, as a design test. The clear implication is that coursework-heavy proposals will struggle to win approval unless they can show integrity against AI assistance. His caution extended to the classroom, where he argued over-reliance on the tools risks “less learning, less satisfaction, less usefulness in the workplace”.

Notably, he was even-handed about AI’s role in the machinery of assessment. It has “enormous potential” to improve efficiency and cut costs in the qualification system, he acknowledged — but on marking specifically it is “not actually as good as you might think”, because it “makes mistakes, it hallucinates, and it reflects biases”, and can drift beyond the rules humans set. That tension — useful for administration, unreliable for judgement — mirrors what schools themselves are grappling with, as AI use in English schools races ahead of national strategy.

Looking forward

The signal to awarding bodies is unambiguous: build assessments that hold up when every student has a capable AI in their pocket, or expect a harder ride through approval. Bauckham’s closing line — “AI will have an important place, but we must remain in charge” — captures the regulator’s stance. For UK educators and edtech suppliers, the practical work now is designing assessment that is credible in an AI-saturated world, not merely policing its misuse.