Marketplace Tech Police use new AI tool that can identify someone without facial features
Jun 2, 2025
In this discussion, James O'Donnell, an AI reporter at MIT Technology Review, delves into the controversial realm of police surveillance technology. He introduces Track, a new AI tool that identifies individuals using non-facial attributes like clothing and body type, aiming to sidestep the pitfalls of traditional facial recognition. The conversation highlights the ethical concerns surrounding privacy, false positives, and the murky landscape of regulations across U.S. states. O'Donnell underscores the need for greater transparency and community engagement in how these technologies are implemented.
AI Snips
Chapters
Transcript
Episode notes
New AI Surveillance Beyond Faces
- Traditional facial recognition uses AI to analyze face patterns but faces scrutiny for false positives and privacy concerns.
- Newer tools skip faces and analyze clothing, hair, and body type, expanding surveillance without needing facial details.
Expanded Surveillance Scale Risks
- TRAC increases surveillance scale by tracking individuals even without clear facial views.
- This expansion risks heightened surveillance and introduces potentially large amounts of false positives.
Demand AI Transparency and Oversight
- Police departments should be transparent about their use of AI surveillance tools like TRAC.
- Communities need opportunities to review and regulate such technologies before adoption.