Parliamentary evidence: 8 September 2026 · England · Reviewed: 14 September 2026
The House of Commons Education Committee heard evidence from assessment bodies about artificial intelligence and educational integrity. This was evidence to an inquiry, not a Government decision or a new regulatory requirement.
AQA reported 96 AI-related malpractice cases across 3.5 million qualifications this year, while stressing the wider burden and uncertainty faced by teachers. The Standards and Testing Agency said local-authority moderators had reported possible AI use in primary writing for the first time, although it had received no direct evidence or maladministration reports. Ofqual’s chief regulator said detection tools are struggling to keep pace with model capability and described an approach of avoiding, adapting and deterring AI-related risks in extended-writing assessment.
For educators
The emerging emphasis is less on whether a piece of text can be labelled “AI-generated” and more on whether a pupil’s learning process can be understood and trusted. Schools can make reasoning visible through an unaided first response, a record of sources and AI contributions, revisions made, confidence expressed and a short oral explanation. This should be treated as a formative teaching approach, not presented as an Ofqual requirement.