92% of UK graduate employers caught candidates misusing AI — a record
The News
Some 92% of UK graduate employers suspected or confirmed student misconduct in assessment and selection this year, up from 85% in 2025 and 78% in 2024, according to the Institute of Student Employers' Student Recruitment Report 2026. Unauthorised AI use in interviews and online assessments is the most common offence, reported by 66% of employers, while 82% now publish guidance on acceptable AI use — up from 55%. Employers are responding with personalised and randomised questions, time limits and automated eligibility screening, and 77% say they are reviewing or redesigning their processes. Graduate vacancies grew 4% this year after an 8% fall, at 141 applications per role.
Why It Matters
The candidate side of the AI arms race has gone mainstream, and it is asymmetric: candidates are automating at industrial scale while only 16% of these employers use AI in their own recruitment. The result is a tide of near-identical applications that slows your screeners and buries genuine candidates. It also corrodes the basic premise of your funnel — that an application represents a real person's real effort — and 77% of employers are now redesigning assessment around that loss of trust.
Taking Action
All action is situated in your corporate context. These are starting points, not prescriptions — choose what fits.
Connect
Who to talk with in your enterprise
Your early-careers and campus recruiting lead, plus whoever owns assessment design. Ask what they are actually seeing on interview integrity, and whether your published AI-use guidance matches what you can realistically police at each stage.
Inspect
What to take a closer look at
Audit your graduate pipeline for the tell-tale signs: applications that read alike, over-formal or stilted interview answers, and sudden spikes in near-identical responses. Then check whether your assessment design — time limits, randomisation, proctoring — is built to detect AI use or simply assumes it away.
Reflect
What to think about more deeply
If 92% of employers suspect candidates of using AI, is the problem the candidates cheating — or the assessment you have designed that a chatbot can pass?