In the Institute of Student Employers' public account of its Student Recruitment Survey 2026, 92% of surveyed employers said they suspected or confirmed student misconduct in assessment and selection. ISE says the most common forms involved unauthorised AI use during interviews and online assessments. The figure measures employers reporting a concern; the public account does not say how many candidates used AI against the rules or how many incidents were proven.

The survey, published on October 14, draws on 155 UK student employers and their 2025/26 recruitment cycle. ISE says those employers handled more than 1.7 million applications. Its public summary reports 31,577 hires, although its report access page describes that total as hires and work experience placements. The employer share suspecting or confirming misconduct rose from 85% in 2025 and 78% in 2024, according to the summary. It does not disclose the full questionnaire or the evidence employers used to classify a case as suspected or confirmed.

The big change

  • What changed: ISE's employer survey shows AI use becoming a more explicitly bounded part of recruitment. Employers increasingly accept help with written applications, while 82% now give acceptable-use guidance and assessments usually restrict AI.
  • Why it matters: An applicant can use AI to present genuine experience in a CV and still breach an employer's rules by using it during an interview or test. Clear stage-specific instructions help candidates know what work must be their own.
  • What to watch: With 77% of surveyed employers reviewing or redesigning recruitment, hiring teams face a concrete choice: which stages permit AI assistance, which test independent ability, and how to explain that distinction before an assessment begins.

Permission depends on the stage

ISE says employer acceptance of AI in the hiring process rose to 64% from 46% the previous year, particularly for help writing application materials. That does not amount to blanket permission throughout a recruitment process. The same public account says candidates are usually not allowed to use AI during online assessments, interviews or assessment centres.

The practical distinction is between assistance that helps present a candidate's own experience and outside help during an exercise meant to assess that candidate's abilities. ISE advises applicants to check each employer's rules and include only skills and achievements they can support. Its head of insights, Claire Tyler, says employers want to understand what a candidate can genuinely bring to the role.

Employers are changing their processes too. ISE says 77% are reviewing or redesigning recruitment, and describes measures including more personal questions, time limits and randomised questions. The summary does not establish which measure works, or whether any change is a response to AI use alone. High application volumes are another pressure in ISE's account: it reports 141 applications per graduate role.

What the 92% figure cannot tell us

The reported share is about employers, not applications or candidates. It does not reveal the number of cases within each employer's recruitment process. The public summary groups suspicion and confirmation together, without giving a breakdown or explaining the detection method. It therefore cannot support a claim that 92% of students cheated, that AI detection is accurate, or that misconduct caused the rise in applications.

The full ISE report requires membership or paid access, and the data dashboards are member-only. BIG CHANGE used the public summary, not the restricted report or dashboard. Question wording, response counts for individual measures and any subgroup analysis are unavailable in the material we could inspect.

BIG CHANGE previously covered HackerRank's Chakra AI interviewer, a vendor product for assessing candidates. ISE's survey describes employers' recruitment practices and concerns across its respondent group.

Sources & further reading