How to spot fake and AI-generated candidates in 2026
Hiring has a new problem. AI tools can now write a flawless CV, a tailored cover letter, and convincing interview answers in seconds, which means a growing share of applications come from people misrepresenting who they are or what they can do. Some are mildly polished; some are outright fake, including organized schemes that apply at scale. For a small team that can't afford a bad hire, knowing how to tell real from generated matters more than ever.
Here's how to spot fake and AI-assisted candidates, and screen them out without insulting the genuine ones.
What you're watching for
There are two different problems, and they need different responses. The first is candidates using AI to polish genuine applications, which is normal and mostly fine; the person is real, they only had help writing. The second is misrepresentation: fabricated experience, fake identities, someone other than the applicant doing the work, or AI generating answers to skills the person doesn't have. You're not trying to ban AI help. You're trying to confirm the person is who they say and can do what they claim.
Signs worth a second look
No single signal proves anything, but a cluster of these warrants closer attention:
Flawless but generic. A CV and cover letter that are polished yet say nothing specific about your company or role often came from a generator. Real enthusiasm references real details.
Answers that don't match the CV. A candidate whose written application is impressive but who can't speak to their own listed experience in conversation is a red flag.
Vague on specifics. When asked to walk through a project they claim to have done, they stay high-level and can't go deep. People who did the work can go deep.
Inconsistencies. Dates, titles, or details that don't line up across their CV, profile, and answers.
Reluctance to turn the camera on, or odd delays in a video interview, which can signal someone other than the applicant is answering.
How to screen without insulting real candidates
The trick is to verify in ways that feel like normal hiring, not interrogation. Genuine candidates breeze through these; fakes struggle.
Ask for specifics, live. A short conversation where you ask someone to walk through a real project, decision by decision, is the single best filter. Generated text can't survive "what happened next, and why did you choose that?"
Give a small, real task. A brief, role-relevant exercise tied to the real job shows you what someone can do, not what they can describe. Keep it short out of respect for their time.
Use targeted screening questions. Ask a couple of questions on the application that need genuine experience to answer well, the kind a generator handles generically and a real practitioner answers with substance.
Talk to them. A real conversation, by phone or video, surfaces most fakes fast. The gap between a polished written application and a vague live answer is the tell.
Don't over-correct
A word of caution: in hunting for fakes, don't start treating every candidate as a suspect. Plenty of strong, genuine people use AI to help write a CV, and plenty are nervous or awkward on camera for honest reasons. Lead with verification through real conversation and real tasks, not accusation. The goal is to confirm the good ones, not to interrogate everyone.
Let your process do the filtering
The defense against fake candidates isn't paranoia; it's a process that surfaces substance. Screening questions on the application, a quick scoring pass to prioritise, and a real conversation early catch most misrepresentation before it costs you time. In KalosHR you can attach screening questions to any role, score applicants against your requirements, and keep every answer and note in one place, so the gap between a slick application and a thin real answer is easy to see.
AI has changed what an application can hide. The answer is the oldest one in hiring: talk to people, ask them to show their work, and trust what they do over what's written down.
