A new study from AQA explores whether artificial intelligence can support the development of questions for national, high-stakes examinations, where speed must be balanced with fairness, reliability and alignment to curriculum standards.
As part of the study, 203 students completed a GCSE-level Biology practice paper containing both AI-authored and human-authored questions. The questions were then evaluated through three different lenses:
By combining student performance data with professional and expert review, the study moves the discussion beyond what AI can generate to consider how those questions perform in practice.
Patrick Coates, CEO of the e-Assessment Association, said: “This study provides valuable real-world evidence about how AI-authored questions perform, rather than judging them simply because AI was used to create them. The findings show genuine potential, while also identifying where human expertise, robust quality assurance and further testing remain essential. We encourage assessment professionals, awarding organisations and regulators to download the report and consider how its evidence can support responsible experimentation and adoption.”
The research contributes valuable evidence to the assessment community as awarding organisations, regulators and assessment providers consider where AI can add genuine value and what safeguards are needed.
Read more and access the full AQA report.
Join the global community advancing e-assessment through innovation, research, and collaboration.
Keep informed
This site uses cookies to monitor site performance and provide a mode responsive and personalised experience. You must agree to our use of certain cookies. For more information on how we use and manage cookies, please read our Privacy Policy.