AI applications force recruiters to rethink candidate verification

Artificial intelligence is changing how candidates apply for jobs and how employers decide whom to interview. Reports by BBC News and The Guardian describe a rise in tools that can improve CVs and cover letters, generate answers to application questions, or suggest responses during online interviews. The shift is forcing recruiters to reconsider what counts as acceptable assistance, how they test skills, and how they verify that a candidate’s performance is genuine.

Assistance at the application stage

Tom, a 22-year-old geography graduate from Maidstone, told BBC News that he applied for more than 200 jobs before securing a restaurant manager role in September 2026. He described the process as exhausting and demoralising. He said AI helped him write CVs and cover letters and identify keywords likely to be noticed by automated screening. He said he did not use it during an assessment or interview.

Tom argued that applying for 10 to 20 jobs a day makes it difficult to produce a completely tailored application each time. He also called it hypocritical for companies to prohibit AI when some employers use AI themselves to screen candidates. His account illustrates the central tension: candidates want to present their strongest case, but excessive automation can make an application less personal or suggest skills the applicant does not possess.

The Institute of Student Employers surveyed 155 of the UK’s largest graduate employers in 2026. The trade body said employers were becoming more accepting of AI at some stages of recruitment. The most commonly accepted uses were preparing CVs and cover letters, followed by answering online application questions. Acceptance, however, does not mean that every use is authorised.

Misconduct reaches a reported high

The ISE said 92% of surveyed employers had encountered or suspected misconduct during selection. That compares with 85% in the previous survey and 78% two years earlier. The Guardian reported that 66% of respondents had seen at least one case of a graduate misusing AI during an interview. The survey did not name the employers reporting those incidents.

The most frequent problem was unauthorised assistance during interviews and online assessments. According to the Guardian’s account of the ISE survey, 47% of respondents reported candidates using AI to complete online assessments. The next categories were AI use for online application questions and for CVs or cover letters, at 10% each. These figures describe employer reports, not independently verified totals for every recruitment process.

Recruiters say real-time assistants can be detectable. Chris Morrow, who runs Manchester-based Digitalent and recruits for the AI sector, told The Guardian that candidates may pause while waiting for a chatbot, use filler phrases, or speak in unusually formal and unnatural language. He compared the result to reading from an autocue. Morrow said his agency normally ends an interview when it identifies such behaviour and considers it cheating.

The pressure behind mass applications

AI also makes it easier to send many applications at once, a practice known in recruitment as “spray and pray”. A trainee accountant who asked BBC News not to use his real name said he had applied for 50 graduate jobs and used AI in every application for research, CV editing and mock interview questions. He said he would not use it to cheat in an assessment, although he knew people who might do so.

Stephen Isherwood, chief executive of the ISE, said the practice was increasing competition and pressure on students. The ISE said graduate roles currently attracted an average of 141 applications each. The market had stabilised after a sharp fall in the previous year, but remained difficult. In the Guardian’s report, the ISE said graduate hiring rose 4% in 2026 after an 8% decline in 2025, and employers expected a further 7% rise over the next 12 months.

The survey also showed different employer experiences. Forty per cent of respondents said they were reducing graduate intake, 45% were increasing it and 15% were maintaining current levels. Isherwood attributed reductions to economic uncertainty rather than AI. He said the larger effect of AI on graduate employment was more likely to emerge in two or three years.

Employers redesign verification

EDF closed its graduate recruitment scheme early after receiving 12,000 applications, a 10% increase on the previous year. Hollie Moody, the company’s head of early careers recruitment, said AI had probably contributed to the increase and was being used more freely during selection. EDF now provides guidance on acceptable use and is trying to make its process more resistant to undisclosed assistance.

EDF reads CVs later in the process and uses skills- and behaviour-based online assessments and video interviews. The company did not currently use AI screening, although Moody said it might have to in future because of application volumes. She described the issue as a moral dilemma and said the company wanted authenticity, personality and diversity of thought.

For organisations, the practical response is not simply to search for supposedly machine-written prose. Recruiters need clear rules explaining when AI may help with formatting, research or preparation and when it becomes unauthorised assistance. They can combine structured interviews, skills exercises, follow-up questions and checks that ask candidates to explain their own work. Such measures do not prove that every suspicious answer involved AI, so allegations should remain carefully attributed and candidates should not be treated as dishonest solely because their language sounds polished.

The change for organisations in English-speaking markets is therefore procedural. Employers must disclose their expectations, test relevant abilities directly and verify that the person interviewed can perform the work without a hidden real-time assistant. Candidates, meanwhile, need to understand the role and present their own experience accurately. AI may support preparation, but undisclosed answers during interviews can undermine both fairness and the reliability of hiring decisions.