The use of AI in internal assessments | NCFE

What can we help you find?

The use of AI in internal assessments

NCFE supports the overall approach taken within this guidance. Learners must submit work which is their own, and that submitting work which is not their own, and/or which shows other artificial intelligence (AI) misuse, will be investigated for potential malpractice and could result in disqualification from an assessment, or qualification overall.

Where AI is used by a learner to support the completion of an assessment, the JCQ guidance explains how this should be acknowledged by the learner, and how markers should ensure that learners are demonstrating independent thinking.

The updated JCQ guidance focuses on some real-life examples of malpractice where AI has not been used appropriately, as well as giving providers information on ways to support with verification of learner work, and some tips on how to approach internal assessing when you suspect or know AI tools have been used. For more specific guidance from NCFE, you can also visit this page.

Internal assessment design

The design of internal assessments can go some way to mitigating the problems that generative AI brings. Here are some tips from our assessment experts on designing assessments that are less susceptible to AI misuse by learners:

1) Break down the assessment into smaller tasks that are submitted separately but form a whole assessment.
2) Use a range of evidence types within a single unit.
3) Use authentic workplace activities such as observations, professional discussions, and presentations as part of your assessment mix.
4) Ask learners to personalise their submissions with examples from their own lived experience.
5) Ask learners to record and submit a learning journal as they complete their assessment.
6) Carry out some elements of the assessment under direct supervision of a teacher.

AI writing detectors

AI writing detectors must never be used as the sole method for identifying AI misuse. All AI writing detectors can deliver false positives and false negatives and there is often little explanation provided as to how the software reached its decision; as such some organisations do not allow their use to determine authenticity. Centres using detection tools for plagiarism and AI misconduct should use these tools as part of a broader strategy for assessing learner assignment authenticity. Teachers, familiar with their learners, are best suited to determine work originality.