The rise of tools such as ChatGPT has rightly prompted concern about how we might continue to assess secondary school students as fairly as possible. The interim report of the curriculum and assessment review notes these concerns, but how should it respond?
Some commentators argue that we must now move more exclusively to nationally standardised, timed, terminal assessments to guarantee reliability at key assessment points.
Daisy Christodoulou, who recently wrote for this publication’s ‘Curriculum Conversation’ feature about assessment reform, has claimed on social media that “AI is making all forms of non-exam assessment totally obsolete”.
I have sympathy with fears over AI’s impact on fairness and rigour in student assessment, yet the more alarmist reactions risk throwing away the baby (non-examined assessment in subjects like art, design technology, drama, and film studies) with the AI bathwater. Doing so would only narrow opportunity.
Non-examined assessment (NEA) refers to coursework or project-based tasks completed over time, often incorporating stages of research, planning, practical execution, and evaluation. It is especially valuable in more typically creative subjects where process and iteration can be as important as outcome.
Until recently, its critics focused on its workload implications. Now, those same critics have seized on the advent of AI to precipitate its downfall – often conveniently ignoring that AI also makes assessment itself (examined or otherwise) quicker and more efficient.
It is true that student work completed outside of the classroom is increasingly likely to be influenced in part or in whole by AI. That is an argument for building more regular, in-class assessments into our teaching practice.
In addition, together with parents, we have a role to play in educating ourselves and our pupils about the risks of cognitive off-loading (letting AI do our hard thinking for us). End-of-unit tests, unseen response tasks, timed essays under controlled conditions and ‘old school’ handwritten terminal exams are part of countering that negative influence.
We can’t safeguard creative learning by eliminating it
But that is more relevant to university assessment reform than to schools. Meanwhile, properly administered NEA encourages difficult thinking as much as unchecked AI use undermines it. Indeed, asserting that NEA is ‘totally obsolete’ is short-sighted.
In truth, carefully structured NEA, particularly when integrated into classroom practice and accompanied by regular formal checkpoints, allows students to apply their learning practically, reinforces process-based thinking and encourages creative engagement.
Christodoulou agrees, qualifying her remarks by saying that it is ‘unsupervised writing assessments [that] are no longer viable’. The emphasis here should be on ‘unsupervised’.
I can only speak to my subject, Film Studies. In my experience NEA has oversight built into the process. For example, the OCR specification explicitly states that “work must be undertaken under direct teacher supervision”.
Teachers must monitor key production stages and authenticate all submitted materials. This is definitely not ‘unsupervised writing’ — but it also isn’t a timed exam. It is scaffolded, stage-based, and pedagogically integrated coursework.
NEA in this context becomes a space where academic and creative knowledge intertwine. Theory is not only something to be remembered and examined in formal conditions, it is also something to be tested and challenged through practice.
Of course, there are other concerns about NEA, not least school-level dishonesty and teacher bias (as well as workload). But we can’t safeguard creative learning by eliminating it, and we can’t preserve humanity (or protect children’s future prospects) in the age of AI by retreating into exam halls.
Instead, let’s take this opportunity to adapt. In Film Studies, this might include viva-style discussions about a student’s production decisions, with peer feedback built into the assessment, as well as honest reflection on where AI was (or wasn’t) used and why, promoting transparency.
In fact, there is a strong case for reinstating the heavier weighting of NEA within the qualification as a whole (as was the case before the last curriculum reforms).
We don’t have to choose between the rigour of the final exam and creative practice. What we need is non-examined assessment that rewards hard thinking and hard work.
Perhaps AI can even help us devise it.
Your thoughts