As the integration of AI has become a staple across multiple fields, the hiring process is now the newest place for organizational security breaches. What began as using AI to polish a resume or to speed up the selection process has now created a digital weak point for companies and candidates alike. AI that can coach both sides of an interview and provide answers in real time continues to evolve. Now, it looks like a program that can fabricate identities, manipulate interviews, and gain direct access to sensitive internal databases.
The FBI's 2025 Internet Crime Report documented $13 million in losses from AI-enabled employment fraud. This included deep-fake technology used to deceive hiring managers during live interviews.1 Adaptive programs being used to perpetuate AI candidate fraud fall into a massive blind spot. Traditional hiring practices weren't made to catch this. The vetting and verification process now must go beyond general screening. Internal defenses must be enhanced so that organizations aren't leaving their most critical systems exposed.
AI-assisted overstatements, falsified credentials, and document manipulation
This wave of candidate-assisted AI employment fraud specifically targets all aspects of the interview process. Real-time deep-fake video filters can alter a candidate’s appearance. Proxy applicant schemes involve one person successfully completing interviews while a different individual performs the job once hired. Fictional identities and fabricated personas can pass an initial AI screening before any human interaction is triggered. But do all of these AI candidate fraud attempts stem from bad actors or nefarious intent? A 2025 survey of 3,000 job seekers found that only 26% of applicants believe AI will evaluate them fairly.2 Numerous examples of AI bias and the poor algorithmic structure of third-party matching software have led to a "systemic rejection" phenomenon.3 This makes AI both the weapon and shield being used by candidates and companies. Whoever has the 'smarter' system will be the victor.
The same survey found that 4 out of 10 applicants already use AI during the hiring process for resumes, cover letters, and assessments. Of those surveyed, 6% admitted to participating in proxy interview fraud, either by posing as someone else or having someone else assist or pose as them.4 This presents a scenario in the next few years where employers have to assume a percentage of the candidate profiles they’re seeing could be fake. A recent hiring study predicts that by 2028, one in four applicants participate in some form of AI-generated resume fraud.5 This means that a well-matched resume is no longer strong evidence that a candidate holds the experience needed for a role. This also means that AI is being used in nearly every piece of this dynamic, whether you are the employer, the candidate, or the threat in question.
Open risks and automated, high-volume defense: The AI double-edged sword
Traditional verification is no longer enough to combat the litany of weak points appearing in the modern hiring process. This is especially reflected in government systems where pre-verification was already high, but legacy technology levels the playing field. Threat actors can weaponize AI across every stage of the hiring funnel. Generative AI can be used to form Applicant Tracking Systems (ATS)-optimized résumés and cover letters. Real-time deep-fake interview video and voice tools can craft appealing on-camera identities through live interviews. AI chat assistance can feed candidates answers during remote assessments without detection. Automated application generation now overwhelms recruiters and screening systems with sheer volume. A single-point verification at hire is no longer sufficient when the person who passed the interview may not be the person logging in on day one.
Experian's 2026 forecast specifically flags generative AI's ability to create synthetic candidates capable of passing video interviews in real time. They warn that employers may unknowingly onboard people who are not who they claim to be.6 This risk bubble isn't centralized either. The U.S. Department of Justice cases document a multi-year scheme using stolen identities from at least 80 U.S. citizens to place workers at more than 100 companies. The scheme generated over $5 million for North Korea before facilitators were sentenced.7 International “laptop farms” can propagate these cyberattacks through long-form, slow-detection methodologies such as credential sharing, siphoning work, and 'after hiring' access distribution.8
However, the same technology creating the threat has become the primary mode of industry defense. Here are some prime examples of how AI can be used to prevent fraud:
- Organizations can deploy AI-powered hiring fraud and identity-analytics platforms. These can be used to flag anomalies across résumé data, device fingerprints, and behavioral signals.
- Third-party video platforms can be optimized with voice-authentication and liveness-detection technology. These optimizations can be designed to specifically catch deep-fake artifacts.
- HR departments can be trained to utilize tools that encourage continuous identity verification. This includes other risk-signal monitoring, programs that track devices, location, network, and behavioral analytics throughout the candidate journey.
- Third-party service providers can now assist with procedural countermeasures such as document and credential verification services; manager education on fraud indicators; formal investigation reporting processes; more in-person interviews; and three-way video handoffs between supplier, manager, and candidate.
This changes the trajectory of the AI arms race. Initially, verification must move from a single checkpoint to continuous assurance across the employment lifecycle. There should be clear agreements between organizations and clients about who owns authentication responsibility at each handoff. Next, participants will need to consent to liveness detection measures that confirm identity and credentials continuously. This would include key engagement points like application, interview, onboarding, day 1, day 30, and beyond. The goal would be to ensure the résumé matches not only the background check, but the person who shows up to work.
Human oversight & continuous identity verification are the best solutions for structural gaps
In state and local government, human oversight and continuous monitoring are still the foundational defenses against hiring fraud. Depending on their role, government employees can access sensitive citizen data, classified information, and critical infrastructure systems. While technology provides detection tools, it's human judgment that catches what automated systems miss. Human oversight ensures accountability at every handoff.
This means government hiring managers must be equipped to recognize fraud indicators in real time. Other organizational changes will likely need to happen in tandem. Policy updates are needed to gather candidate consent for cameras being kept on during interviews. This also means that the verification process cannot end once the hiring paperwork is approved. Post-assignment monitoring at 30, 60, 90+ days should become critical for roles with system access. This kind of monitoring would incorporate:
- biometric and geographic checks (in addition to MFA) at login
- surprise video calls to verify the worker matches the hired identity
- insider risk protocols that validate device IP addresses live on video
While these changes in workflows and hiring might be interpreted as overcorrection, the proliferation of candidate fraud (thanks to AI) requires a different tact. In government and highly sensitive roles, increased job stakes demand a "trust but validate" culture, where controls are proportionate to the risks. Audits on both sides should be required. A third-party validator should check that an organization's verification processes work, and that their hiring managers are properly trained and held accountable for maintaining vigilance.
Verification depth: Building a process that catches AI candidate fraud before an offer
The challenge hiring teams are faced with now is protecting their organization without sacrificing the candidate experience. The best verification workflows have embedded technology that qualified candidates barely notice, while a fraudster on the other end is deterred before they gain access. A combination of identity verification, adaptive staffing solutions, embedded security processes, and updated training are all ways to defend against threats.
This can be complex, but with the right service partner, you will be able to manage a cohesive and modern verification infrastructure accordingly. This will ensure that your workforce is not only vetted and legitimate but also equipped to meet evolving compliance and operational demands.
Let CAI help your organization with end-to-end support, staffing, and security adaptation, so your company has what it needs to stay protected and competitive.
Endnotes
- FBI National Press Office. "Cryptocurrency and AI Scams Bilk Americans of Billions" FBI.gov April 06 2026. https://www.fbi.gov/news/press-releases/cryptocurrency-and-ai-scams-bilk-americans-of-billions. ↩
- Mary Baker. "AI Fuels Mistrust Between Employers and Job Candidates" Gartner. July 31 2025. https://www.gartner.com/en/newsroom/press-releases/2025-07-31-gartner-survey-shows-just-26-percent-of-job-applicants-trust-ai-will-fairly-evaluate-them. ↩
- Stefan Milne. "People mirror AI systems’ hiring biases, study finds" UW News. November 10, 2025. https://www.washington.edu/news/2025/11/10/people-mirror-ai-systems-hiring-biases-study-finds/. ↩
- Baker, "AI Fuels Mistrust." https://www.gartner.com/en/newsroom/press-releases/2025-07-31-gartner-survey-shows-just-26-percent-of-job-applicants-trust-ai-will-fairly-evaluate-them. ↩
- Taylor Nicole Rogers. "Why 1 in 4 job applications could be fake by 2028" the LA Times. July 30th 2026. https://www.latimes.com/business/story/2026-07-30/why-one-in-four-job-applications-could-be-fake-by-2028. ↩
- Kathleen Petters. "Experian’s Future Of Fraud Forecast: Why Fighting Fraud Matters More Than Ever" Experian. January 13, 2026. https://www.experian.com/blogs/news/2026/01/13/experians-future-of-fraud-forecast-2026/. ↩
- SIA Editorial Staff. "Fraud scheme: US sanctions North Korean firm and individuals" SIA. March 18 2026. https://www.staffingindustry.com/editorial/cws-30-contingent-workforce-strategies/fraud-scheme-us-sanctions-north-korean-firm-and-individuals. ↩
- Patrick J. McMahon, Mark J. Neuberger. "Laptop Farms Highlight Identity Fraud Risks Of Remote Work" Foley. July 13th 2026. https://www.foley.com/insights/publications/2026/07/laptop-farms-highlight-identity-fraud-risks-of-remote-work/. ↩