Veritone’s “Track” AI system uses body data to sidestep facial recognition bans

A controversial AI tool is helping law enforcement circumvent facial recognition bans across the U.S. by tracking individuals through alternative physical characteristics. This technology raises significant privacy concerns as it expands to federal agencies during a period of increased surveillance, potentially creating a new frontier in public monitoring that operates in legal gray areas where facial recognition has been restricted.

How it works: Veritone‘s “Track” AI system identifies people using non-facial attributes like body size, gender, hair characteristics, clothing, and accessories rather than biometric facial data.

• The system can create timelines tracking individuals across different locations and video feeds, even when faces are obscured or not visible.
• Currently operating only on recorded video, Veritone plans to expand the technology to analyze live video feeds in the future.

The big picture: Veritone CEO Ryan Steelberg explicitly designed Track to circumvent facial recognition restrictions while still enabling surveillance capabilities.

• “If we’re not allowed to track people’s faces, how do we assist in trying to potentially identify criminals or malicious behavior or activity?” Steelberg explained about the vision behind the product.
• The system is available through major cloud platforms including Amazon and Microsoft.

Widespread adoption: Track is already being used by 400 customers across the United States, including state and local police departments and universities.

• The Department of Justice began using Track for criminal investigations in August 2024.
• Veritone’s broader suite of AI tools, which includes traditional facial recognition technology, is also used by the Department of Homeland Security and Department of Defense.

Privacy concerns: The ACLU identified Track as the first non-biometric tracking system deployed at scale in the United States, warning it creates novel privacy challenges.

• Civil liberties experts characterize the technology as “potentially authoritarian” in its surveillance capabilities.
• The expansion occurs as the Trump administration pushes federal agencies to increase monitoring of protesters, immigrants, and students.

Surveillance inputs: The system can analyze footage from various sources to track individuals across environments.

• Potential video sources include police body cameras, drones, public videos on platforms like YouTube, and citizen-submitted footage from Ring cameras and cell phones.