AI marking tools show bias in research. Nobody is sure whose problem it is to fix.
Studies at UK universities have found that automated marking systems can disadvantage students from certain backgrounds. Developers say the models can be improved. Regulators have not yet decided what to require.
By Wistl Editorial · · Labs
AI-assisted marking and feedback tools have been positioned as one of the most commercially significant applications of AI in education. The government allocated £1 million in February specifically to accelerate AI-powered marking and feedback tools. The government's AI in Education roadmap situates this within a broader commitment to reducing teacher workload through technology. Research conducted at UK universities into how these tools perform across different student populations has produced concerning findings. The British Journal of Educational Technology published research in 2025 from t