Thousands of UK Students Caught Cheating With AI as Detection Lags Behind
LONDON, UK — June 18, 2025 A Guardian investigation reveals that nearly 7,000 confirmed cases of AI-enabled cheating were recorded at UK universities during the 2023–24 academic year—an uptick from 1.6 per 1,000 students in 2022–23 to 5.1 per 1,000. The figure is expected to rise to 7.5 per 1,000, though experts stress these are likely only the tip of the iceberg.
Meanwhile, traditional plagiarism has dropped significantly, falling from 19 to 15.2 cases per 1,000 in 2023–24, with projections showing a further decline to around 8.5 per 1,000. Yet tracking of AI-related misconduct remains inconsistent: over 27% of universities don’t record it separately , and detection tools struggle with sophisticated AI and paraphrasing software.
Experts warn detection is becoming a cat-and-mouse game. A study from the University of Reading found 94% of AI-generated submissions went undetected. Dr Peter Scarfe commented that even when detectors flag AI use, "it is near impossible to prove," noting the challenge of prosecuting suspects without risk .
Students view AI as a productivity booster. Many report using it to brainstorm, summarize, or structure work—particularly those with learning difficulties—though a minority admit to submitting AI-generated text without edits .
Universities are under pressure to update assessments and integrity policies. Academics call for moves toward skill-based tests and mandatory supervision. The UK government is injecting over £187 million into AI literacy and assessment reform, urging institutions to harness AI’s benefits while curbing misuse.