The problem HR is facing
Recruiters in 2026 describe two connected problems. First, application floods: when AI writes CVs and cover letters in seconds, every opening gets more, and more similar, applications. Second, candidate fraud: people who are not who they say they are.
- In a Gartner survey of 3,000 job candidates in 2025, 4 in 10 said they use AI during the application process, mostly for CVs, cover letters and assessment answers.
- 6% admitted to interview fraud: posing as someone else, or having someone else stand in for them.
- Gartner predicts that by 2028, 1 in 4 candidate profiles worldwide could be fake. That is a forecast, not a measurement.
Using AI to write a better CV is not fraud. Faking who you are is. The challenge for HR is to stop the second without punishing the first.
How many fake candidates reach your interviews?
A simple hiring funnel. Change the numbers to match one of your roles. The catch rate of a check step is your own assumption.
A way forward
The tempting answer is more AI screening on top of AI applications. Gartner's own advice points elsewhere: layers, and humans at the right moment.
- Say what AI use is fine. Tell candidates openly: using AI to polish your CV is fine, having someone else do your interview is not. Clear rules reduce grey zone cheating and are fairer to honest candidates.
- Put one live, human moment early. A short video call with the camera on, or one in person step for final rounds. In Gartner's research, 62% of candidates said they were more likely to apply when roles required an in person interview.
- Verify identity at the offer stage, with the same check for everyone. Checking only some candidates invites bias.
- Use structured interviews. The same job related questions for everyone, scored the same way, are harder to fake with a script and predict job performance better (Schmidt and Hunter, 1998).
What this means for HR
- Fairness cuts both ways. Every extra check also burdens honest candidates. Keep checks short, the same for everyone, and explained.
- Remote roles need the most care. News reports of fake candidates mostly involve fully remote jobs, so that is where one in person step matters most.
- The law is watching. In the EU, AI used to select candidates is high risk under the AI Act, with the main duties applying from 2 December 2027. Tools that judge emotions from faces or voices are already banned at work.
Honest limits
The 6% is self reported, so the real share may be higher. Most other fraud figures circulating in 2026 come from companies that sell detection tools. The share of fake applicants is your own estimate, because no reliable measurement exists. The simulator assumes fakes are spread evenly through the funnel: good screening would lower the risk, AI polished applications that look better than real ones would raise it.
Sources
- Gartner (2025). Candidate survey and prediction on fake candidate profiles, reported by HR Dive, 8 August 2025. source
- Schmidt, F. L., & Hunter, J. E. (1998). The validity and utility of selection methods in personnel psychology. Psychological Bulletin, 124(2), 262 to 274.