UK DfE Pilots ChatGPT Scripts for KS2 Writing Moderation to Cut Assessment Costs
The UK government is testing AI-generated writing samples to train KS2 moderators as a cost-reduction step, but this risks decoupling calibration from real student error patterns. No peer-reviewed validation of moderator accuracy under the new regime has been released. The move raises immediate questions about the integrity of national writing standards used to evaluate millions of primary pupils.
Stakeholders including the National Education Union have requested sight of the pilot data and an independent review by Ofqual before the change is locked into the 2025 assessment cycle. Absent transparent methodology and outcome metrics, the cost-saving measure may quietly erode the very literacy benchmarks it is meant to safeguard.
Ofqual: By September 2025, moderator agreement on authentic KS2 scripts will fall below 82 percent in any local authority using exclusively AI-generated training materials.
Sources (3)
- [1]Primary Source(https://www.newscientist.com/article/2584776-test-moderators-use-ai-generated-writing-to-judge-literacy-standards/)
- [2]Supporting Source(https://www.gov.uk/government/publications/key-stage-2-tests-writing-moderation)
- [3]Supporting Source(https://doi.org/10.1038/s41562-023-01692-3)