online job interviews are becoming a growing concern for employers. What once seemed like a marginal issue is now moving rapidly into mainstream recruitment practice.
With tools such as ChatGPT capable of generating polished answers in real time, recruiters may increasingly find themselves wondering whether they are speaking to the candidate or to the candidate’s AI assistant.
When traditional interview signals become easier to manufacture, employers need to redesign the way they assess talent.
AI is weakening traditional hiring signals
Generative AI is making it easier for candidates to produce strong CVs, cover letters and interview answers. That does not automatically mean candidates are unqualified, but it does make traditional selection signals less reliable.
A polished answer in a remote interview no longer necessarily proves that a candidate can think clearly, solve problems independently or communicate effectively under pressure.
For employers, the challenge is not simply to detect AI use. The real challenge is to design interviews that reveal how candidates think, decide and respond in situations where generic AI-generated answers are less useful.
Do not rely on AI detection alone
Trying to catch candidates in the act may seem tempting, but detection-based approaches are risky. Asking candidates to share their screen or monitor their behaviour too closely can damage trust and create a poor candidate experience.
It may also exclude good candidates unfairly. Small delays, looking away from the camera or using notes do not automatically prove that someone is being assisted by AI.
Instead, employers should focus on interview design. A well-structured interview makes it harder for candidates to rely on generic AI output and easier for recruiters to assess genuine competence.
Ask candidates to evaluate flawed material
One practical approach is to give candidates imperfect material and ask them to critique it. This could be a flawed business plan, a weak market analysis, an incorrect code snippet or an unrealistic project proposal.
This type of exercise shifts the assessment from polished presentation to critical thinking. The candidate must explain what is wrong, why it matters and how they would improve it.
AI can still support candidates, but it becomes much harder to hide behind generic responses when the task requires judgement, prioritisation and domain knowledge.
Use follow-up questions more deliberately
Follow-up questions are one of the strongest tools recruiters have. When a candidate gives a strong answer, the interviewer can ask for a specific example, a trade-off, a counterargument or a moment when the same approach failed.
This forces the conversation into more personal and contextual territory. Candidates who truly understand their own experience can usually explain the details. Candidates relying heavily on AI-generated answers often struggle when the discussion moves beyond the first polished response.
The goal is not to ban AI from recruitment, but to make sure interviews still measure the human capabilities that matter.
Test real-time judgement
Employers can also introduce short live exercises during the interview. These do not need to be complex. A recruiter might ask a candidate to prioritise three competing tasks, respond to a realistic client situation or explain how they would handle an unexpected problem.
The point is to observe reasoning in the moment. How does the candidate ask clarifying questions? How do they structure their thinking? Do they notice missing information? Can they make a decision without over-relying on a perfect script?
These behaviours are difficult to fake and often reveal more than a standard competency answer.
Be transparent about AI use
Employers should also clarify their position on AI use in the recruitment process. Some organisations may allow candidates to use AI for preparation, but not during live interviews. Others may actively encourage AI use for certain assignments if the role itself requires AI literacy.
What matters most is consistency. Candidates should know what is allowed, what is not allowed and why.
A clear AI policy helps avoid confusion and creates a fairer process for everyone involved.
Recruitment needs a new assessment mindset
The rise of AI-assisted candidates does not mean online interviews are obsolete. But it does mean employers can no longer rely on the same signals in the same way.
Recruiters need to move from scripted questions to stronger assessment design. That means more realistic tasks, deeper follow-up questions, clearer expectations and a sharper focus on judgement, adaptability and communication.
AI is changing how candidates prepare and perform. The employers that respond best will not be those that try hardest to police every tool, but those that redesign hiring around the capabilities AI cannot easily replace.
