Deepfake job applicants are showing up in hiring interviews, report finds

A new iprospectcheck survey of 1,500 U.S. hiring managers and owners found rising concern about AI, identity fraud and FCRA compliance in background screening. One in four employers said they have already interviewed a suspected deepfake or proxy candidate, even as most still want humans making final hiring and screening decisions.

Why it matters: - Employers are dealing with a hiring environment where identity fraud, AI-generated applications and compliance risk are converging. - The findings suggest more companies may need tighter identity checks, better screening documentation and clearer human review steps. - The report also shows employers are worried about fast-moving fraud tactics before those tactics become routine.

What happened: - iprospectcheck surveyed 1,500 U.S. business managers and owners in July 2026 for The 2026 State of Screening Report. - The survey examined how employers are using AI in hiring, handling candidate fraud, managing background screening and meeting FCRA requirements. - The company released the report on August 5, 2026.

The details: - 22.9% of employers said they may have unknowingly interviewed a proxy or deepfake candidate. - 18.5% said they have directly encountered or suspected candidate fraud. - 29.3% said they have hired someone who later did not seem to be the same person they had previously interviewed. - 5.1% of companies said they do not formally verify identity before hiring. - 22.3% said they are not sure what their identity verification process covers. - 80.9% said identity verification should be a standard part of background screening. - 67.5% said identity verification should apply to every role, not just sensitive ones. - 31.6% of companies said they use AI anywhere in hiring. - 15.8% said they use AI to review resumes. - 13.8% said AI has significantly improved hiring. - 25.0% said AI has made no real difference. - 61.2% said a human should always review AI hiring recommendations. - 73.4% said a human should review every potentially adverse background check record before it is reported. - 12.5% said they would trust AI to make hiring recommendations unsupervised. - 43.7% of companies said they reviewed disclosure and authorization forms in the past year. - 12.6% said they do not use those forms at all. - 35.5% of respondents named accuracy as the most important factor when choosing a screening provider. - 18.1% named price as the top factor. - 83.8% said U.S.-based support is very or somewhat important. - 46.5% said they would pay more for U.S.-based support. - 42.2% expect flat screening budgets over the next 12 months. - 28.6% expect budgets to increase. - 17.5% said they are investing in AI-assisted screening. - 14.9% said they are investing in identity verification. - 22.7% named AI-generated applications as the biggest screening risk over the next five years. - The report said full findings and methodology are available at The 2026 State of Screening Report.

Between the lines: - The report shows employers want AI tools, but not AI-only decisions. - The emphasis on human review suggests hiring teams see automation as support, not a replacement for accountability. - The gap between confidence and documentation points to compliance risk if employers face an adverse action challenge. - Accuracy and U.S.-based support outranked price, which suggests screening buyers are prioritizing risk reduction over cost alone.

What's next: - Employers are likely to keep adding identity verification and AI-assisted screening as they respond to fraud risks. - The report points to more demand for human oversight in hiring workflows and background screening decisions. - iprospectcheck said its U.S.-based verification and compliance teams continue to support employers across healthcare, transportation, financial services, staffing, warehousing, retail and professional services.

The bottom line: - Deepfake and proxy candidates are no longer a theoretical threat in hiring. - Employers appear ready to use AI, but only with humans still making the final call.