The Institute of Student Employers' 2026 annual survey delivers a sharp datapoint: 92% of its 290 member employers — firms like Marks & Spencer, Rolls-Royce, and Accenture — now suspect or confirm candidate misconduct during hiring. That's up from 85% the prior year and 78% the year before, a trajectory that suggests not a blip but a structural shift in how graduates approach the job market. The dominant vector is AI-assisted interview cheating. Two-thirds of respondents (66%) reported at least one incident of a graduate using AI tools during a live interview. The mechanics are crude but effective: apps that listen to interviewer questions in real time and generate suggested answers, visible to the candidate on a second screen during video calls. Chris Morrow, who runs Manchester-based recruitment agency Digitalent and hires for the AI sector itself, says the tells are obvious — filler phrases to buy processing time, unnaturally formal language, a visible reading cadence. His firm ends the interview on the spot. Beyond live interviews, 47% of employers caught AI use in online assessments, and 10% flagged AI-generated application materials. The gradient matters: interview cheating is adversarial and real-time, a fundamentally different category from polishing a cover letter with ChatGPT. The ISE survey lumps them together, but the risk profiles are not the same. One is tool-assisted drafting; the other is impersonation. ISE chief executive Stephen Isherwood frames the problem as self-defeating for candidates: "If you misrepresent yourself, you could end up doing the wrong job." That's true but incomplete. The systemic cost is a degradation of the signal that interviews are supposed to produce. If employers cannot trust that the person answering is the person they'd be hiring, the entire selection mechanism loses information value. The labor market context complicates the narrative. Graduate hiring fell 8% last year but recovered 4% in 2026, with a further 7% rise expected. Crucially, Isherwood attributes any hiring contraction to macroeconomic uncertainty, not AI displacement — a notable pushback against the prevailing doom narrative. Forty percent of firms cut their graduate intake while 45% expanded it, suggesting firm-level variation driven by sector economics rather than a technology-wide hiring freeze. The gap in the story is the employer response. The survey documents rising misconduct but says almost nothing about countermeasures. Are firms shifting to in-person interviews? Redesigning assessments to be AI-resistant? Accepting that AI-assisted answers are the new baseline and evaluating differently? The absence of any systematic defensive adaptation, despite three consecutive years of rising misconduct, suggests the hiring industry is documenting a problem it has not yet decided how to solve. Isherwood's estimate that AI's real impact on graduate jobs is "two or three years down the line" deserves scrutiny. If 92% of employers already see misconduct and 66% catch interview cheating, the impact is not forthcoming — it is here. The question is whether the hiring system adapts its filters or simply becomes noisier, rewarding candidates who are best at gaming AI tools rather than those best suited for the work.
