Home / Education / AI Sparks Debate in UK School Writing Assessments
AI Sparks Debate in UK School Writing Assessments
14 Aug
Summary
- UK DfE tests AI for pupil writing samples, saving 95% costs.
- Concerns raised over AI samples lacking authentic pupil voice.
- AI use could disadvantage non-native or neurodivergent students.
The UK Department for Education (DfE) is experimenting with AI, specifically GPT-5, to generate writing samples for assessing primary to secondary school pupil literacy standards. This pilot program aims to drastically cut annual costs by 95%, from around £100,000 to a much lower figure.
This year and next, some moderators will review AI-generated text instead of actual pupils' work for standardization. The DfE plans to decide by spring 2027 whether to continue this AI-driven approach or revert to external suppliers for samples.
However, the move has sparked ethical debates. Experts highlight concerns that AI samples might not represent authentic student writing, potentially shifting perceptions of acceptable literacy standards. There are also worries that AI's tendency towards bland writing could disadvantage English as a second language speakers or neurodivergent students.
DfE acknowledges these risks, noting LLMs can exclude atypical language. They plan mitigation through thorough review and editing of AI outputs. Educators also fear a 'backwash' effect, where AI-generated materials inadvertently become the new teaching standard.
