JCQ sets the rules for authenticating coursework and non-exam assessment. Here's what its AI guidance requires in plain English, and what it means for your department in practice.
What JCQ guidance covers
The Joint Council for Qualifications sets the rules for how non-exam assessment (NEA) and coursework are authenticated across the main UK exam boards. Its guidance on AI addresses a specific question: how do you confirm that work submitted for a qualification is genuinely the candidate's own when AI tools are freely available? Always check the current version on the JCQ website, as it is updated.
The core principle
Work submitted for a qualification must be the candidate's own. Where AI has been used, it must be acknowledged, and any AI-generated content that a candidate presents as their own constitutes malpractice. That principle is not new — it's the long-standing rule about authenticity applied to a new tool.
The practical implication for teachers: your signature on an authentication declaration is a professional statement that you believe the work is the candidate's own. That responsibility sits with you and your school, which is why supervision and knowing your pupils' work matter far more than any software.
What it means in practice
Acknowledgement of AI use
Where a candidate has used AI, it must be declared and referenced, with the prompts and outputs recorded as appropriate to the subject. Undeclared use presented as the candidate's own is malpractice.
Teacher authentication
Teachers must be confident work is the candidate's own before signing. Where you cannot be confident, you should not sign — and your school's malpractice process applies.
Supervision is the strongest safeguard
The guidance points towards supervised conditions for key parts of NEA. In practice this is the most reliable route: work produced under supervision needs no forensic investigation.
Keeping evidence of process
Plans, drafts and progressive versions demonstrate authenticity far better than a finished piece alone.
What it does not say
It does not require or endorse the use of AI detection software as a means of authentication. As we cover in our piece on AI detectors, that software is unreliable and biased, and a detector score would not constitute evidence in a malpractice case.
What schools should do
- Make sure all staff involved in NEA understand the authentication requirement
- Increase supervised time for key assessed work where practical
- Keep drafts and process evidence routinely
- Tell candidates clearly what is and isn't permitted, in writing, before they start
- Have a clear internal route when a teacher cannot authenticate work
- Reference the requirements in your school AI policy
Tell pupils early and clearly
Much malpractice at this level stems from genuine confusion rather than intent to cheat. Candidates often don't realise that using AI to "help improve" a paragraph crosses a line. Explicit, written, subject-specific guidance before they start prevents most problems — and is far kinder than dealing with it afterwards. See teaching pupils to use AI thoughtfully.
Keep checking the current version
This area is being actively updated by JCQ and the boards. Rather than relying on a summary — including this one — build an annual check of the current JCQ position into your exams and assessment planning.
Need staff clear on assessment and AI?
We help schools translate the guidance into practice staff understand. AskColin gives schools low-cost monthly support with one-to-one help whenever you need it — so when you're stuck on a real job at 8am, there's someone to ask. Practical, jargon-free, built around your department.
Get one-to-one AI support for your schoolFrequently asked questions
What does JCQ say about AI in assessments?
Work submitted for a qualification must be the candidate's own. Where AI has been used it must be acknowledged and referenced, and AI-generated content presented as the candidate's own work constitutes malpractice. Teachers must be confident work is authentic before signing authentication declarations.
Does JCQ guidance require AI detection software?
No. The guidance does not require or endorse detection software as a means of authentication. Detectors are unreliable and biased, and a detector score would not constitute evidence in a malpractice case. Supervision and process evidence are the recommended safeguards.
How can schools authenticate NEA work with AI available?
Increase supervised time for key assessed work, keep drafts and progressive versions as process evidence, ensure staff understand the authentication requirement, and tell candidates clearly in writing what is and isn't permitted before they start.
What if a teacher can't authenticate a pupil's work?
They should not sign the declaration, and the school's malpractice process applies. The signature is a professional statement that you believe the work is the candidate's own, so schools need a clear internal route for when that confidence isn't there.