DHS Added Deepfake Detection to Identity Rally
The Department of Homeland Security expanded its document validation testing to address generative AI threats.
Updated on Sept. 21, 2026 in Cybersecurity
Live Poll
Do you trust government testing to effectively stop the use of AI-generated fake identities?
The Department of Homeland Security Science and Technology Directorate has expanded its Remote Identity Validation Rally (RIVR) to include a dedicated challenge for deepfake and AI-generated ID detection. This program update aims to measure performance and risk against evolving identity fraud threats.
Why it matters
As generative AI lowers the barrier for sophisticated identity fraud, the DHS is pivoting its testing program to quantify how well commercial identity systems can resist these specific digital attacks. This initiative bridges the gap between current identity-proofing standards and real-world attack vectors.
Previous 2025 testing revealed that systems faced a median false-match rate increase of 11 times, while one system, DVS 6, recorded a 3.15 percent false rejection rate and a 0.88 percent false acceptance rate.
The players
Department of Homeland Security Science and Technology Directorate
The research and development arm of the federal agency that manages testing and evaluation for identity verification standards.
National Institute of Standards and Technology
The federal laboratory that develops technical standards, including the Special Publication 800-63A-4 guidelines for identity proofing.
Homeland Security Investigations
The investigative component of U.S. Immigration and Customs Enforcement that maintains a laboratory for analyzing fraudulent document methodologies.
The details
The DHS evaluates identity systems by testing them against both genuine and fraudulent document datasets, including those sourced from the Homeland Security Investigations laboratory. By introducing a dedicated deepfake challenge, the agency shifts from static document matching to analyzing whether biometric and presentation attack detection tools can distinguish between authentic identities and AI-synthesized replicas. These tests provide a performance benchmark against mandated thresholds, such as a 10 percent error maximum and a more stringent 1 percent goal.
Timeline
July 31, 2025: NIST finalized federal identity-proofing guidelines.
September 30, 2026: The DHS hosts an informational webinar for prospective technology providers.
October 2026: Applications open for document validation testing.
November 2026: The DHS begins soliciting identity deepfake attack examples from external organizations.
December 2026: The DHS begins accepting systems for deepfake evaluation.
The Tech Race
This effort aligns commercial system performance with the rigorous NIST Special Publication 800-63A-4 federal identity-proofing guidelines. It represents a shift from foundational document checks to an arms race against generative AI capabilities that bypass traditional validation.
This program determines which identity verification systems are cleared for federal use, eventually influencing the security standards used in consumer financial and government services. Providers aiming to supply these sectors must now meet the new deepfake detection benchmarks starting in December 2026.
The takeaway
The DHS is codifying performance benchmarks for deepfake detection to stabilize identity verification in a generative AI environment. Stakeholders should monitor the outcomes of the 2026 rally to see which vendors meet the 1 percent high-performance error goal.
What happens next
The DHS informational webinar is scheduled for September 30, 2026, followed by the opening of application windows for document validation testing in October 2026.
Further reading
For broader context on federal security standards, see Cybersecurity.
Live Poll
Do you trust government testing to effectively stop the use of AI-generated fake identities?
