DHS expands RIVR testing to deepfakes and AI-generated IDs

The Department of Homeland Security is expanding its RIVR 2026 evaluation to include a dedicated biometric deepfake detection category, covering AI-generated impersonations in video chats. Applications for document validation open in October, with PAD and deepfake systems accepted from December; DHS will also solicit deepfake attack samples starting in November.

• RIVR 2026 adds a dedicated biometric deepfake detection category
• Document validation applications open in October; PAD and deepfake systems in December
• From November DHS will collect deepfake attack samples and creation methods
• No system met the 1% error goal in the 2025 RIVR