The Deepfake Candidate: Building Identity Checks into Remote Hiring

Key takeaways
- Rising AI usage: A survey of 3,290 candidates found that 39% reported using AI during the application process (shrm.org).
- Interview fraud prevalence: Among 3,000 candidates surveyed, 6% admitted to committing some form of interview fraud (shrm.org).
- Identity verification gaps: While 76.8% of teams verify candidate identity at some stage, only 53% do so before an offer is made (pin.com).
- Future projections: By 2028, it is expected that one in four candidates globally could be fake (shrm.org).
A survey of 3,290 job candidates found that 39% reported using AI during the application process (shrm.org). This data highlights a significant shift in how applicants approach the initial stages of recruitment. While many use these tools for efficiency, the rise of generative technology introduces new complexities for identity verification and the assessment of authentic candidate skill sets.
The spectrum of AI usage in applications
To understand the risk to hiring integrity, we must distinguish between assistance and deception. The data shows that AI usage is already widespread across different stages of the application lifecycle. Among candidates who use AI, half used it to generate cover letters, 36% used it for writing samples, and 29% used it to draft assessment answers (shrm.org).
The mechanism behind this behavior is often a desire for efficiency or a perceived need to compete in a high-volume market. When a candidate uses AI to draft a cover letter, they are using a tool to polish their presentation. When they use it for a writing sample or an assessment, they are using it to generate the core substance of their response. This creates a distinction between "augmentation," where a candidate uses AI to improve their own work, and "substitution," where the AI performs the task entirely.
What the data says is that AI is a standard part of the modern application toolkit (shrm.org). What it does not say is whether these candidates are being transparent about their use of these tools. For HR teams, the challenge is not the use of AI itself, but the potential for this usage to mask a candidate's true capabilities or, in more extreme cases, to facilitate identity fraud.
Defining the scope of candidate fraud
Candidate fraud is defined as any deliberate deception by an applicant, which includes fabricated work history, stolen or synthetic identities, paid stand-ins, or deepfakes (pin.com). This definition covers a broad range of behaviors, from minor exaggerations on a CV to the use of sophisticated digital personas designed to deceive a remote interviewer.
The scale of this issue is reflected in recent surveys. A separate survey of 3,000 job candidates found that 6% admitted to committing some form of interview fraud, such as posing as someone else or having someone else pose as them (shrm.org). This 6% represents a specific subset of behavior: the act of impersonation. This is a distinct escalation from using AI to write a letter; it is a direct attempt to bypass the fundamental requirement of a job interview, which is the assessment of the person being hired.
The distinction between using AI to assist a task and using AI to impersonate a person is the new frontier for identity verification in remote hiring.
Why It Matters
The mechanics of identity deception
As remote hiring becomes the standard, the "digital interface" has become the primary way we meet candidates. This creates a vulnerability. If a candidate uses a deepfake to mimic the appearance or voice of a qualified individual during a video interview, the traditional visual and auditory cues used by recruiters may no longer be sufficient to guarantee identity.
Analysis: The mechanism of deepfake fraud relies on the lag between the emergence of the technology and the adoption of verification protocols. In a traditional office setting, identity is verified through physical presence and government-issued ID presented in person. In a remote setting, identity is verified through a screen. If the screen can be manipulated, the entire interview process is compromised.
There is a counter-reading to this concern. Some argue that the prevalence of AI tools in the hands of all candidates levels the playing field. If every candidate uses AI to write their assessments, the "authenticity" of the writing sample becomes a moot point, and recruiters must shift their focus entirely to real-time verbal communication or live technical testing.
| Category of AI/Fraud Usage | Reported Prevalence (Survey Data) |
|---|---|
| General AI use in applications | 39% (shrm.org) |
| AI for cover letters (among AI users) | 50% (shrm.org) |
| AI for writing samples (among AI users) | 36% (shrm.org) |
| AI for assessment answers (among AI users) | 29% (shrm.org) |
| Interview fraud/impersonation (total sample) | 6% (shrm.org) |
Verification timing and the "Offer Gap"
Current practices show a significant gap in when identity is actually checked. In a survey conducted by Pin, 76.8% of teams verify candidate identity at some stage, which suggests that most organizations recognize the need for some level of validation (pin.com). However, only 53% of those teams perform this check before the offer is made (pin.com).
This creates a "verification gap" in the middle of the hiring funnel. When identity checks are delayed until after an offer is extended, the cost of fraud increases. If a company discovers an identity mismatch after the contract is signed, they have already invested significant time and resources into onboarding and background checks. The goal for HR is to move verification earlier in the process to prevent these sunk costs.
Analysis: Comparing the 76.8% of teams that verify identity at some stage against the 53% that do so before an offer suggests that many organizations are using "reactive" verification. They check identity as part of a final background check, rather than as a "proactive" step during the interview stage. Proactive verification is a preventative measure; reactive verification is a detection measure used after the risk has already entered the system.
Layered verification strategies
To manage the risk of deepfakes and synthetic identities, HR teams should consider a layered approach. Relying on a single point of verification is no longer sufficient in a remote-first environment. A layered approach involves multiple, independent checks that are difficult to spoof simultaneously.
One layer is identity verification through government-issued documentation via a third-party service. Another layer is the use of live, unscripted technical assessments. A third layer is the use of "liveness" checks during video interviews, which require the candidate to perform specific movements to prove they are a real person and not a digital projection.
Identity verification should be viewed as a continuous process throughout the hiring lifecycle, rather than a single checkbox at the end of the process.
Watch For
What HR can do on Monday morning
We cannot wait for technology to solve the problem of deepfakes; we must adjust our processes to account for the technology that already exists. If the expectation is that one in four candidates globally could be fake by 2028 (shrm.org), then the current protocols are already outdated.
To prepare, HR teams should implement these three steps:
- Audit the verification window: Review your current hiring workflow to identify exactly at which stage identity is verified. Aim to move the verification step to occur before the verbal offer is extended.
- Update interview protocols: Incorporate "liveness" checks or unexpected, unscripted questions into video interviews to disrupt the ability of a candidate to use a sophisticated digital overlay.
- Standardize AI disclosure: Update your application instructions to clearly state the organization's policy on AI usage for assessments and writing samples. This establishes a baseline for what constitutes "assistance" versus "deception."
The landscape of recruitment is changing. As the distinction between human and synthetic identity blurs, the role of HR shifts from simply finding talent to verifying the authenticity of the talent being found. By moving toward proactive, layered verification, we can protect the integrity of our teams and the fairness of our hiring processes.