JCQ AI Guidance for NEA and Coursework: A Practical UK Schools Guide

By Tomas Lerpiniere · Updated 2026-10-28

Joint Council for Qualifications (JCQ) expects centres to protect assessment authenticity while teaching learners how generative AI may and may not be used. This guide translates that into a workable workflow for GCSE, A level NEA, EPQ, BTEC and similar internal assessments — without treating detector percentages as proof.

In this guide

Feedback Flows tools in this guide

AI writing detector Writing fingerprints Reference checker Similarity Scoring

What JCQ expects of centres

JCQ’s AI Use in Assessments materials are guidance for centres, not a product endorsement list. In practice, schools and colleges need: a clear local AI policy for coursework, staff training on what to look for, learner education on acknowledgement, and a human-led process when work looks inconsistent with known performance.

Detector tools — including Feedback Flows’ free AI writing detector and the full AI writing detector — sit inside that process as evidence prompts. They do not replace professional judgement or awarding-organisation rules.

Pre- vs post-authentication stages

Timing matters. If concerns arise before the learner signs the declaration of authentication, most centres can resolve the issue under internal policy (conversation, redraft under supervised conditions, support notes). Once the declaration is signed, suspected malpractice usually follows JCQ reporting expectations for that qualification and series.

Staff should diary when the declaration was signed, keep contemporaneous notes, and avoid relying on a single tool output if the case may leave the centre. Pair this article with our school AI policy guide so SLT and subject leads share the same language.

Acceptable help vs unacknowledged AI

Scenario Typical centre approach
Brainstorming ideas or explaining a concept with AI, then writing in the learner’s own words with acknowledgement where policy requires Often permitted if local policy allows and the final text is authentic
Using AI to generate paragraphs submitted as original NEA / coursework without acknowledgement Treat as potential malpractice; gather process evidence
Teacher-provided PEEL/PETAL frames and modelled sentences Pedagogy — not automatic AI misuse; interpret detector scores carefully
Fabricated or mismatched references in a bibliography Investigate with citation checks; may coincide with generative drafting

Evidence that stands up in a centre investigation

  1. Quoted passages. Prefer tools that highlight exact sentences you can show the learner — try the free AI detector.
  2. Writing history. Compare with earlier authentic samples via writing fingerprints or the free authorship checker.
  3. References. Screen bibliographies with the free citation checker for hallucinated DOIs.
  4. Process pack. Plans, drafts, source notes, and a short viva record of how the learner explained their argument.
  5. Recorded outcome. Note discussion held, redraft agreed, or escalation — never “failed by software”.

For why lone percentages fail as evidence, read evidence over AI detection scores. For a full authorship workflow, see writing fingerprints and authorship checks in UK schools.

A fair four-step classroom workflow

  1. Set writer context (EAL / younger writer) and run a paste check on the suspect section.
  2. Read flagged coverage and verified quotes — not only the headline percentage.
  3. Hold a calm 5–10 minute viva; ask for drafts and source explanation.
  4. Decide under centre policy: support and redraft, further checks, or formal reporting after authentication.

Linguistic tells and a ready check-in script are covered in how teachers spot ChatGPT in student essays.

FAQ

No. JCQ’s AI Use in Assessments guidance distinguishes permitted assistance (with acknowledgement where required) from unacknowledged AI generation submitted as the learner’s own work. Centres must set clear local rules and teach learners what to declare.

Timing depends on whether the learner has signed the declaration of authentication. Concerns before authentication are usually handled under centre policy; after signing, suspected malpractice typically follows JCQ reporting routes. Always check the current JCQ booklet for your series.

No. Detector scores are assistive signals. JCQ expects centres to gather holistic evidence and exercise human professional judgement. Prefer quoted passages, draft history, authorship comparisons and viva notes over a raw percentage.

Keep the flagged submission with highlighted quotes, earlier drafts or planning notes, any writing-fingerprint comparison, citation/DOI checks, and a brief record of the learner conversation and outcome.

Explain early what help is allowed, what must be acknowledged, and that process evidence (plans, drafts, source notes) may be reviewed. Frame detectors as conversation starters, not automatic fails.

Use the free AI writing detector at Feedback Flows to paste a section, set EAL or younger-writer context, and review sentence-level highlights before deciding whether a formal check is needed.

Continue exploring

Feedback Flows tools for this workflow

Integrity AI writing detector A hybrid check with writer context for EAL and younger writers, highlighted passages you can quote exactly, and a place to record what you d...
Integrity Writing fingerprints Build a style profile from previous authentic samples, then compare a new submission. See how vocabulary, structure and habits align — evide...
Integrity Reference checker Check the whole submission — not just the reference list. Match in-text citations to the bibliography, verify bibliographic details, and opt...
Integrity Similarity Scoring Every marked rubric and automarking submission can enter your organisation's private similarity bank. See a similarity %, flag work at 50% o...