Universities are facing increasing challenges in addressing the use of artificial intelligence (AI) by students, with some academics warning that institutions may be conceding defeat in efforts to prevent AI-assisted cheating. Katherine Rundell, an award-winning author and University of Oxford fellow, expressed concern that many universities have effectively raised a “white flag” as students increasingly use large language models to complete academic work.

Speaking at the Hay Festival, Rundell highlighted that AI tools now enable students to produce essays on books they have not read, contributing to what she described as a “vast counterfeiting of knowledge.” She noted that while some educators believe they can detect AI-generated work, it is primarily those who use the technology poorly who are caught. “When they are using it well,” Rundell said, “and then moving with it over three or four hours, we can’t tell.”

Some universities have begun allowing AI use provided it is properly acknowledged, a practice Rundell criticized as a dereliction of duty. She argued that this approach undermines the fundamental academic expectation that students develop their own critical thinking skills. Rundell cautioned that it must not become possible to complete humanities degrees without engaging with the source material.

Data from Russell Group universities reveals a sharp increase in disciplinary actions related to AI misuse. More than 2,000 undergraduates were sanctioned for improper use of generative AI tools during the 2024-25 academic year, a significant rise from around 700 cases in the previous year. Students cited low contact hours and unsatisfactory teaching quality as drivers for turning to AI, particularly for research and note-taking tasks.

Several recent graduates provided insight into their AI usage. A neurodivergent law master’s student noted that AI had become essential for managing academic work, especially given language barriers and the level of support available for disabilities. Another master’s student with dyscalculia said they used AI to compensate for inadequate university resources, helping to level the academic playing field.

A final-year history student described using AI assistants like ChatGPT and Claude primarily to refine ideas rather than generate original analysis, while observing that peers in business management relied heavily on AI to complete essays with minimal personal effort. “Some have done as little as two hours of essay work all year,” the student said, noting that AI could efficiently locate sources and help secure strong grades.

Nick Hillman, director of the Higher Education Policy Institute, acknowledged the complex position universities face. He suggested that AI has the potential to enhance the capabilities of otherwise capable students by streamlining reasonable tasks. However, Hillman also linked the proliferation of AI use to student frustration over poor feedback and grading standards, which may be encouraging more widespread reliance on such technologies.

As AI tools become more sophisticated and accessible, institutions grapple with balancing innovation and academic integrity, amid debate over how best to incorporate or regulate these technologies within higher education.