Deepfake candidates now top the list of hiring worries, but the reflex to bolt on more screening tests can drive away the honest applicants you want most. The fix is a more human hiring process, not a more suspicious one. Here is how to do both at once.
One of my clients recently asked me for advice on a growing hiring concern. They discovered a candidate who was perfect in every way, on paper at least. However, while conducting the virtual interview, some strange behavior emerged. The video was glitching, the candidate would ramble for a few minutes, and then proceeded to give a flawless answer, which almost seemed too perfect. Then this happened with a second candidate, and then a third.
Unfortunately, it is not only a possibility but something we are seeing become more and more common. Deepfakes and other forms of AI interview fraud are on the rise, and talent leaders now rank fraudulent or AI-assisted candidates as their single biggest recruitment challenge. Failing to spot these cheats could not only mean that your company hires someone unqualified, but also cost you at least 30% of that employee's first-year earnings, a floor that only climbs once you factor in lost productivity and low morale. Even worse, that fake candidate could be the first phase of a larger fraud operation targeting your company.
While this is a serious issue and can sound quite alarming, I advise my clients not to panic as the right process improvements can greatly reduce the organization's risk. It is important to balance these measures carefully, ensuring that genuine applicants have a great candidate experience while keeping bad actors at bay. The answer is to build an intentional hiring process that focuses on making the experience more human than ever.
Know what deepfake candidates and AI hiring fraud look like
You can't defend against this if you can't identify it. Fraud has been a growing concern for hiring managers, which has continued to escalate as technology advances. Candidates have historically submitted false resumes, provided fake references, or fabricated qualifications. Rigorous screening can usually prevent these candidates from getting too far into the hiring process.
AI deepfakes, however, present a whole new type of threat. Some deepfakes are almost impossible to spot at first glance, and recruiters have learned to start watching out for certain red flags, such as:
- Unnatural pauses and flatness. Real candidates vary their pace during interviews. They may speak slowly to allow time to think, or speed up when they're excited or on familiar ground. A fraudulent candidate may talk with the same rhythm throughout, and may pause shortly before each answer, with the same cadence in their answers.
- A story that drifts. Frauds often have trouble keeping their story straight, even when AI is helping them. This might not be obvious at first, but over time you may notice growing inconsistencies in their story.
- Answers without substance or clarity. If a candidate is getting answers from AI, interviewers might notice they're saying a lot of words, but not offering a lot of detail. Follow-up questions can really fluster these candidates.
- Video that glitches. Deepfake filters still struggle with micro-expressions and sudden movement, so the image can distort when a candidate turns their head or moves a hand across their face.
Gartner predicts that one in four candidate profiles worldwide could be fake by 2028, which turns screening for authenticity into part of every hiring process rather than a special case. If you want the fuller picture of how these scams work, our earlier guide to fake job applicants still holds up.
However, it's important to remember that the vast majority of candidates are not using AI to defraud employers. In fact, many genuine applicants are finding the application process more exhausting than ever... because of AI.
Understand how AI is changing the hiring experience
AI can be a legitimate part of the hiring process. Many employers now rely on AI tools for screening, communications, and organization, and these tools can have real benefits.
However, there's also a downside to AI in hiring, especially if you don't think about how AI impacts the candidate experience. Job applicants are getting increasingly frustrated:
- Only 26% of job candidates trust that AI will evaluate them fairly, so most walk in already wary.
- 38% of U.S. candidates have withdrawn from a hiring process because it included an AI interview.
- 36% of candidates have turned down an offer after a negative interview experience, which means a clumsy process can cost you the hire at the finish line.
It's important to understand this landscape before making any changes to your hiring process. If candidates are already exhausted by AI, they're not going to enjoy performing additional anti-deepfake tests. You could drive away genuine applicants, and you might not even prevent fraud.
How to build a pro-human, anti-fraud hiring process
The solution is to find a balance: streamline legitimate uses of AI, block fraudulent AI, and retain the human touch in the hiring process. Here's how to find that balance:
1. Be transparent about AI usage
AI is still very new, and there's no universal standard for appropriate AI usage. Nearly four in ten candidates already use AI somewhere in their application, but there's not much clarity about what is and is not acceptable. You can help candidates by giving clear guidelines on when it's okay to use generative AI during the application process, and when use of such tools is forbidden.
You can build a sense of trust by disclosing your company's usage of AI in hiring. Let candidates know in advance when AI will be involved in screening, video interviews, and candidate communications. Try to explain why AI is part of your process, and reassure them that human beings make all of the hiring decisions.
2. Ask consistent interview questions
A consistent set of questions is harder to manipulate than a free-flowing chat. When the interview is structured, a fraudster has much less room to steer the conversation onto rehearsed ground. Your interview team can then use follow-up questions to get more detail when needed, and explore any inconsistencies that arise.
Structured interviews can also benefit genuine candidates. Asking everyone the same questions, and scoring their answers the same way, has been shown to reduce bias and improve candidate selection. It also eases the pressure an honest candidate feels to reach for an edge with AI, because the interview is measuring what they can do rather than the personality they bring to the room.
3. Create a coherent flow throughout the hiring process
Modern hiring is complex, and can involve multiple handovers between recruiters and interviewers. That gets frustrating for candidates, who often have to repeat their whole background every time they meet someone new. It also makes life easier for fraudsters, because inconsistencies in their story might just look like the result of a bad handover.
You can close those gaps without adding to anyone's workload. Give your team one shared place to record what each conversation turned up, so the next interviewer walks in already knowing the story and builds on it instead of starting from scratch. The candidate feels that right away, because nobody makes them repeat their whole background at the next stage, and a strong applicant does not quietly fall through the cracks between one person and the next. It also gives you a single view of the whole process, so the details that do not line up have nowhere to hide.
4. Let tools flag risk, and let people decide
Should deepfake detection be the responsibility of people or anti-fraud tools? There are issues with both approaches. People spot a deepfake only about half the time, which is close to a coin flip, so a gut feeling is a weak reason to turn someone away. The tools have their own blind spot: they misfire more often on some faces than others, with an even worse bias for some demographic groups.
Cooperation is the best approach. If you have anti-deepfake tools, use them to flag concerns. Let your team review the analysis and make the ultimate decision about whether to take further action. This gives you the best chance of a correct decision, and also makes your hiring process more defensible from a compliance perspective.
5. Lean on experienced recruiters
Veteran recruiters know how to identify red flags, simply because they've seen it all before. AI deepfake threats are very new, but an experienced recruiter can pick up on some of the subtle signs that indicate if a candidate is on the level. An experienced person can also help ensure that your hiring process remains human-focused and delivers a great experience for genuine candidates.
If you don't have access to an experienced recruitment expert, you can look at a consultancy engagement. Recruitment Process Outsourcing consultants can help your in-house team build stronger processes that deliver safe, reliable long-term hiring outcomes.
Build the process before you need it
None of this asks you to choose between catching fraud and treating people well. A hiring process that is structured, open about its rules, and careful to write down its decisions does both jobs at once, and it does them from the same effort. You raise the cost of faking, and you make honest candidates feel respected, without running two separate programs to get there.
That kind of deliberate design is also what builds confidence. Employers with a finalized AI governance framework are more than four times as confident in managing AI risk as those without one, yet only 12% have finalized one. The leaders who plan their approach deliberately are the ones pulling ahead.
If you want help building that process, Helios can support you at whatever stage you are in:
- Recruitment services for organizations that need to fill roles now without lowering their guard.
- Talent acquisition consulting for organizations that want to redesign their screening end to end.
- Recruitment process outsourcing for organizations that want a structured process run for them.
- AI consulting for organizations building the AI governance framework behind it all.
Book a call with a Helios HR consultant to design a hiring process that protects your company and respects the people applying.
Frequently asked questions
What is a deepfake candidate?
A deepfake candidate uses AI to fake their identity or performance in an interview, often through a manipulated video feed or answers fed live from a chatbot. The goal ranges from landing a job they aren't qualified for to opening a larger fraud.
How can you spot an AI-assisted candidate during an interview?
Watch for a flat, unchanging speaking rhythm, answers that sound polished but thin on detail, a story that shifts between stages, and video that distorts on sudden movement. None is proof on its own, so treat each as a prompt to ask a sharper follow-up.
Are AI interviews driving away real candidates?
Often, yes. 38% of U.S. candidates have withdrawn from a hiring process because it used an AI interview, and most applicants doubt AI will judge them fairly. Piling on anti-fraud tests without explanation risks losing the people you want to keep.
Should you rely on deepfake detection software?
Use it to flag risk, not to make the call. Detection tools miss manipulated media a fair share of the time and misfire more on some faces than others. Let the software surface concerns, then have an experienced person review and decide.
How do structured interviews reduce hiring fraud?
A consistent set of questions gives a fraudster far less room to steer the conversation onto rehearsed ground. It also lets your team score answers fairly, follow up on gaps, and catch the inconsistencies that fabricated stories tend to produce.
Can you fight hiring fraud without hurting the candidate experience?
Yes, and the same moves do both jobs. A process that is structured, open about where AI is used, and careful to record decisions raises the cost of faking while making honest candidates feel respected. You don't need two separate programs.