AI-driven exam cheating and surveillance disrupt academic integrity in schools
French high school students are using AI tools like ChatGPT and Snapchat AI to cheat during exams, with the French Education Ministry reporting a 30% increase in suspected cheating cases among baccalaureate candidates. In the UK, students have reportedly trained an AI model to predict exam questions with over 90% accuracy to guide their revision.
What happened
In French high schools, students are using AI tools including ChatGPT, Claude, and Snapchat's AI to cheat during in-class exams by photographing test questions and sending them to AI for answers. At a press conference on May 19, French Education Minister Édouard Geffray reported that during the most recent baccalaureate session for final-year high schoolers, cheating cases increased by more than 30%, with 1,208 suspected cases among 1.2 million candidates.
In the UK, a group of secondary school students trained an AI model using two decades of GCSE and A-level exam papers, course content, and classroom information to predict exam questions. They claim their model predicted exam papers across all subjects with 85 per cent accuracy last year and above 90 per cent this year, and report using these predictions to revise for upcoming exams.
Reported impact
Teachers report that supervision has become a game of cat-and-mouse, as phones are easy to hide and students use wireless earbuds, smartwatches, and smart glasses. One student noted that cheating via AI is unfair to those who study seriously and make the effort to work.
Index analysis
Teachers are struggling to identify cheating as phones are easily concealed and students employ multiple devices and methods. Schools and universities have guidance on penalties for cheating but face difficulty policing the use of AI to predict exam questions, which operates in a grey area.
Governance controls
Cell phones are already banned in French middle schools and are set to be prohibited in all high schools beginning the next school year. Oxford University has issued an ethical guide to students stating that AI tools must be used responsibly, ethically, and with integrity, transparency, and critical approach.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.