Recruiters Spot Spike in Real-Time AI Cheating During UK Job Interviews
A survey of UK employers reveals a record high in interview misconduct, with applicants caught using real-time AI tools during virtual screenings.
Job applicants across the UK are increasingly turning to generative artificial intelligence to assist them in real-time during remote screenings, but hiring teams are quickly catching on. According to recent survey data from the Institute of Student Employers (ISE), incidents of misconduct during graduate application and interview processes have hit an all-time high, driven largely by the unapproved use of AI systems.
The ISE, which represents 290 major employers such as Rolls-Royce, Marks & Spencer, and Accenture, found in its annual survey that 92 percent of surveyed organisations suspected or identified improper behaviour during candidate assessments. That figure marks a steady escalation from 85 percent in the previous year and 78 percent the year prior. Unapproved AI assistance during evaluations emerged as the leading grievance, with 66 percent of employers documenting at least one occurrence where an applicant misused AI while interviewing.
Beyond live video conversations, AI misuse has spread across multiple screening stages. Approximately 47 percent of surveyed organisations reported applicants using AI to complete online assessments. Other instances involved candidates leaning on AI to answer application questionnaires or generate CVs and cover letters, the latter noted by 10 percent of respondents.
The telltale signs of real-time AI assistance
While software designed to monitor questions and generate live prompts has proliferated, experienced recruiters note that the technology leaves clear operational footprints. Chris Morrow, founder of the Manchester-based recruitment firm Digitalent, recruits directly for the artificial intelligence industry and notes that assisted candidates quickly stand out.
Morrow explained that candidates using these platforms often look as though they are reading directly from an Autocue. Instead of answering naturally, applicants generate recognizable latency while the underlying model processes the interviewer's speech and streams a response. To bridge these awkward silences, applicants often rely on stalling tactics such as repeating filler phrases like "That’s a good question." Morrow added that the resulting responses tend to sound noticeably unnatural and overly formal. When his firm detects real-time AI generation during video calls, the team treats it as cheating and terminates the screening.
Stephen Isherwood, chief executive of the ISE, highlighted that relying on automated answers carries long-term consequences for job seekers. Misrepresenting personal competence or expertise can push candidates into positions they are unprepared to handle, leading to poor outcomes for both parties. Isherwood emphasized that hiring managers are looking for candidates who demonstrate genuine research and a clear understanding of the roles they target.
Hiring trends and the timeline of AI automation
The ISE's 2026 data also contextualized the broader graduate recruitment environment, downplaying the idea that artificial intelligence is immediately replacing early-career openings. The survey reported that graduate recruitment stabilized with a 4 percent uptick this year after declining 8 percent previously, with businesses forecasting an additional 7 percent expansion over the coming 12 months. Hiring patterns remain split across industries: 45 percent of employers plan to expand their graduate intake, 40 percent plan cuts, and 15 percent intend to keep hiring levels unchanged.
Addressing whether task automation has caused corporate reductions, Isherwood explained that current reductions stem from economic uncertainty and a challenging operating climate rather than companies replacing human staff with software. He noted that the direct workforce impact of AI automation on graduate roles is likely two to three years down the line.
What it means for developers
For software engineers and technical candidates, these findings signal a rapid shift in how technical screenings and behavioral interviews are conducted. As hiring teams become proficient at spotting real-time latency, unnatural cadence, and scripted responses, candidates attempting to rely on hidden prompting tools risk immediate disqualification. Instead, interviewers are refocusing evaluations on organic problem-solving, architectural reasoning, and spontaneous technical explanations that cannot be cleanly delegated to a live chatbot.
At the same time, the surge in AI-assisted tooling shows how deeply large language models and speech-to-text systems are permeating everyday productivity workflows. For developers building the next generation of voice agents, contextual evaluation platforms, or automated interview tooling, benchmarking model latency and linguistic naturalness remains critical. Developers experimenting with conversational applications can test top AI models cheaply through one API at https://apixoai.online, evaluating models from providers like Anthropic, OpenAI, and DeepSeek side by side to balance responsiveness and generation quality.
Source: Use of AI in job interviews on the rise – but UK recruiters say they can spot the cheats — The Guardian AI. Written by the Apixo team from that report.
One key for Claude, GPT, GLM, DeepSeek and more. Pay per token with crypto.
Get your API key
