Are We Interviewing a Person or ChatGPT? What CMI's Hiring Poll Means for Leaders

"Are we interviewing a person or ChatGPT?" That is how Petra Wilton, Director of Policy at the Chartered Management Institute (CMI), framed the findings of a poll CMI published on 2 November 2026. It is a fair summary of a feeling many hiring managers will recognise: the candidate on the screen is fluent, polished and well prepared, and you are no longer sure whose fluency you are hearing.

What the research found

CMI's Managers Voice pulse poll asked 928 practising UK managers about AI in recruitment between 14 and 23 September 2026. The headline findings:

• 58% of hiring managers have noticed or suspected a candidate using AI during an online job interview.
• 90% say candidates use AI to help write CVs or cover letters, up from 79% in November 2024.
• 98% identified at least one negative reaction to AI reliance during interviews. 56% say it feels inappropriate, 50% worry the candidate may be misrepresenting their skills or experience, and 50% say it makes them less confident in the candidate's overall ability.
• 46% say suspected AI use in a virtual interview raises questions about whether a candidate can work independently, and 20% regard it as dishonest.

CMI suggests many employers will go back to in-person interviews to get a true picture of the person.

Then the poll turns the question round. 92% of managers say AI should support rather than replace human judgement. 91% say candidates should be told when AI is being used by an employer in the recruitment process, and 91% say employers should regularly check that their AI systems produce fair outcomes for different candidates.

A separate signal points the same way. Acora, an IT services firm, published an analysis of UK search data on 2 November showing searches for AI "humaniser" tools, which rewrite machine text to read as human, reached 476,119 in the past year, up 94% on two years earlier. Acora says much of that probably comes from students, so treat it as an indication, not a measure.

What it means for leaders

The easy reading is that candidates are cheating and employers need better detection. Esm thinks that misses the more useful point. The discomfort runs in both directions. Managers want to know who is really answering. Candidates want to know who, or what, is really judging them. Both are asking for the same thing: to see where the machine is.

John Sansbury's The Journey to Service Excellence makes this point about service desks, long before interviews were part of the conversation: "Good practice is to advise the user that they are dealing with a chatbot rather than try to disguise the fact as people generally recognise their increasing use, lower cost and effectiveness at handling simple tasks." Esm applies his principle more widely. Where AI sits between two people, the honest default is to say so.

Ian Hughes puts the trust cost plainly in What Stays the Same: "A client will usually forgive you for using a machine. They will struggle to forgive you for lying about it." A candidate is a future colleague. The first thing they learn about how your organisation uses AI is how it treats them.

There is also a design question hiding inside the detection worry. If an interview can be passed by reading out a fluent answer, it may have been testing recall and polish rather than judgement all along. AI has simply made that visible.

How the AI Enablement Toolkit helps

The AI Enablement Toolkit for Leaders is a leadership method, not software. Several of its 26 tools apply directly to hiring.

• Tool 5, the value case canvas, starts with the outcome. What is this interview for? If the answer is "to see how this person thinks under real conditions", design for that: a real problem from the role, talked through, with follow-up questions that no prepared script could anticipate.
• Tool 9, work inventory and opportunity triage, separates what AI may usefully support (scheduling, note-taking, first-pass sorting) from what should stay KEEP HUMAN, because the human act itself creates trust and accountability.
• Tool 10, human and AI role design, makes the line explicit. In screening, AI might ASSIST or RECOMMEND. The decision to reject or hire is HUMAN-LED, with a named person who owns the consequence.
• Tool 12, the trust-by-design check, asks whose convenience a change primarily serves. An AI screening step that saves the employer time but leaves candidates guessing scores badly on low self-orientation.
• Tool 16, the minimum governance check, includes "Users know when and how to challenge or correct the output." For hiring, the users include candidates.

Where training fits

Hiring managers need confidence to design assessments that reveal judgement, to ask candidates openly how they use AI, and to talk about acceptable use without sounding accusatory. Many roles now need people who use AI well, so the goal is not to screen it out but to see it used responsibly. That is the kind of practical, scenario-based work Esm's training and enablement sessions are built around.

Honest limits

CMI's figures are a pulse poll of managers' perceptions. "Noticed or suspected" is not the same as confirmed AI use, and the poll does not tell us how often managers were right.

The toolkit does not include an AI-detection tool, and this article does not recommend one. It is a set of management controls, not legal advice. AI in recruitment touches employment law, equality duties and data protection, including rules on automated decision-making. Check the current rules with a specialist before changing your process. Do not rely on memory.

The question to carry

Before your next round of interviews, ask one thing of your own process: what would a candidate feel misled to discover later? Fix that first. The rest of the conversation about AI in hiring gets easier once both sides of the table can see where the machine is.

Tags: recruitment, AI in hiring, trust, transparency, leadership

Sources

• https://www.managers.org.uk/about-cmi/media-centre/press-releases/ai-use-in-online-interviews-surges-as-british-bosses-fear-theyre-talking-to-robots-new-cmi-report-finds/

• https://acora.com/insights/businesses-are-hitting-pause-on-ai-but-their-employees-arent/