In a recent piece for The Conversation, Jonathan S. Weissman, a principal lecturer of cybersecurity at the Rochester Institute of Technology, explains why a stolen face is a uniquely lasting problem. He opens with a woman walking into a grocery store, her face scanned by a security camera before she even reaches the produce aisle. A stolen password can be reset and a credit card cancelled, but as Weissman puts it, no one can reset or revoke the appearance of their cheekbones.
Below are key highlights from Weissman’s The Guardian article:
- How the systems work. Weissman explains that facial recognition does not store actual photos. It converts a face into a mathematical template that maps the positions and proportions of facial features, then matches later scans against that template. Templates are more secure than raw photos but can still be stolen.
- The risk is not theoretical. Weissman adds that biometric data has already shown up in real breaches, including a 2024 hack of a facial recognition system used at Australian bars and clubs and a 2019 incident at a U.S. Customs and Border Protection subcontractor. Whether stolen biometric data has been exploited remains unclear.
- Faces differ from other biometrics. Weissman notes that fingerprints and iris scans usually require deliberate contact with a scanner. Cameras in public spaces capture faces from a distance without the person’s awareness, and stolen templates can be matched against surveillance footage or online photos to track people.
- Devices versus databases. Weissman explains that modern Apple and many Android phones store biometric data in a dedicated local hardware chip, so a cloud breach does not expose those templates. Centralized corporate databases carry far higher risk, especially when companies rely on third-party vendors.
- A face becomes a primary key. Weissman further adds that finked across databases, a facial template can connect emails, financial records, and home addresses into a “super profile.” Combined with deepfakes or 3D face models, it can also help criminals impersonate someone in systems that require a live face.
- What can be done. Organizations should keep only what they need, encrypt templates, use liveness detection, and follow privacy-by-design principles, adds Weissman. People in California, Illinois, and the European Union can submit data access requests, ask what biometric data a company holds, and in some cases request deletion.
Weissman reminds readers that the convenience of breezing through a security check or store entry comes with a permanent risk to privacy, and that society is still figuring out how to live with widespread biometric collection.
Acknowledgement: This summary was generated using AI writing tools as part of my AI Explorations process. For more information, please refer to the original article listed below.