Jobseekers are being warned against using AI tools to secretly feed them answers during live interviews, as viral videos make it increasingly easy to believe the technology can “beat” modern recruitment systems.
Recent TikTok clips promoting real-time AI interview assistants have attracted millions of views by showing candidates receiving ready-made answers on screen while speaking to an interviewer.
One clip demonstrating an AI assistant during a mock video interview has been viewed more than 3.3 million times, while another showing a candidate secretly receiving answers during an interview has passed 3 million views. The technology can listen to a question and produce a polished response almost instantly for the candidate to read back.
At the same time, viral videos have exposed weaknesses on the other side of the recruitment process. One recent clip showed an automated recruiter apparently continuing an interview despite being fed deliberately nonsensical answers, while WIRED documented a jobseeker allowing ChatGPT Voice to conduct an interview with an AI recruiter on his behalf. Neither experiment resulted in the candidate progressing to a human conversation.
Chris DeCarolis, Senior PEO Advisor at PEO Metrics, says the clips risk creating the impression that fooling the first stage of an interview means a candidate has successfully beaten the hiring process.
Getting through the conversation is not the same as passing
“The dangerous part of these videos is that they make the interview itself look like the entire test. If AI gives you a technically impressive answer and the person or bot on the other side accepts it, it can feel as though you have beaten the system.
“But recruitment rarely ends with one convincing answer. You may still face follow-up questions, a practical assessment, another interview, reference checks or a conversation with somebody who actually works in the role.
“That is where the problem begins. If the first interview creates an impression of knowledge you do not actually have, every later stage becomes an opportunity for that gap to appear.”
Real-time interview assistants can now listen to questions and generate responses while the interview is taking place, allowing candidates to read suggested answers from another screen or overlay.
There have already been cases of recruiters apparently identifying candidates using such tools.
In one widely reported example this summer, a candidate interviewing for a machine-learning role was asked to share his screen after the recruiter became suspicious. The interview was ended after the recruiter concluded that AI assistance was being used.
Some employers can disqualify candidates
Jobseekers should also check an employer’s recruitment rules before assuming AI assistance is acceptable simply because the company itself uses AI.
The British Council, for example, explicitly allows candidates to use AI for activities such as practising interview responses, but says applicants must not use AI-generated scripts or real-time AI assistance during interviews and assessments. Its published policy says candidates can be disqualified if AI is used in ways that breach those rules.
“There is a big difference between using AI to prepare and using it to impersonate your ability in real time. Getting AI to create practice questions, asking it to challenge your answers or using it to identify where your examples are weak can be useful.
“Having it secretly answer the actual interview questions for you is completely different. If an employer has specifically prohibited that, you could be putting the entire application at risk.”
The irony is that employers themselves are increasingly using AI during recruitment.
The Information Commissioner’s Office said earlier this year that 70% of employers expect to increase their use of AI and automation in recruitment over the next five years, with technology already being used for CV screening, online assessments and other stages of hiring.
A 2026 candidate survey by recruitment platform Greenhouse also found AI interviews becoming increasingly common, although only 13% of UK candidates who had completed one said they were clearly told beforehand that AI would be used.
A bot continuing does not mean you passed
DeCarolis says this is especially important when candidates are interviewed by an automated recruiter. He explained: “A bot continuing the conversation does not necessarily mean it thinks you are an outstanding candidate.
“It may simply have been programmed to ask a sequence of questions, capture your answers and pass the information into another part of the recruitment process.
“That is why videos showing somebody saying something ridiculous to an AI recruiter and apparently getting away with it need to be viewed carefully.
“The system moving to the next question is not the same thing as receiving a job offer.”
Government guidance on responsible AI recruitment notes that employers may use automated screening tools to score and rank applications, AI-powered assessments to evaluate skills and recruitment chatbots to interact with candidates.
This means a candidate can appear to have successfully completed an AI interview while still being filtered out later by another automated assessment or a human review.
The follow-up question is where the hack can unravel
DeCarolis says the biggest weakness with real-time AI assistance appears when an interviewer asks a candidate to go deeper.
“AI is very good at producing a confident first answer. The problem is the second and third question. An interviewer might ask why you chose that approach, what went wrong the last time you used it, what you would change or exactly what happened afterwards.
“If the original answer did not come from your own knowledge or experience, maintaining that conversation becomes much harder.”
The same applies to behavioural interview questions.
“Imagine AI gives you the perfect answer about resolving a conflict at work. The interviewer then asks who was involved, exactly what you said and what happened afterwards.
“Suddenly you are no longer reading a polished paragraph. You are being asked to describe an experience that needs to remain consistent and believable.”
Use AI before the interview, not during it
Rather than avoiding AI entirely, DeCarolis recommends using it as a preparation tool.
“Give it the job description and ask it to create ten difficult interview questions. Answer them yourself, then ask the AI to point out where your examples lack detail.
“You can also ask it to challenge you with follow-up questions so that you learn to think through your answers rather than memorise them.
“That gives you an advantage without creating a situation where you are dependent on AI during the actual interview.”
“Jobseekers are understandably frustrated when companies automate large parts of recruitment and then expect candidates to remain entirely human throughout the process.
“But turning it into bot versus bot does not solve that problem. You are ultimately being hired to do the job, not to produce the best AI-assisted conversation. If the employer eventually closes the chatbot, removes the second screen and asks you to demonstrate what you can actually do, you still need to have the answer.”
