Marketplace All-in-One Police use new AI tool that can identify someone without facial features
Jun 2, 2025
James O'Donnell, an AI reporter at MIT Technology Review, dives into the controversial world of AI in policing. He discusses a new tool, Track, which identifies individuals by analyzing body characteristics rather than faces, raising concerns about potential false positives and misidentifications. O'Donnell sheds light on the lack of federal regulations governing these technologies and dives into the ethical implications, including cases of wrongful arrests linked to AI decision-making. The conversation balances innovation with the pressing need for accountability and transparency in law enforcement.
AI Snips
Chapters
Transcript
Episode notes
AI Identifies People Without Faces
- Police use Veritone's TRAC AI tool that identifies people without faces by analyzing clothing, body type, and hair.
- This broadens surveillance from facial recognition to tracking individuals across various video feeds even when faces are obscured.
Expanded Surveillance and False Positives
- TRAC increases the scale of surveillance beyond facial recognition by allowing tracking of people even when faces aren't visible.
- It may also lead to more false positives as many can share similar clothing or body features.
Involve Communities in Surveillance Decisions
- Communities should be consulted before police and tech companies deploy facial recognition or AI surveillance tools.
- Permissions, restrictions, and transparency about tool usage should be established with community input.