Amazon is rolling out AI-powered facial recognition for Ring doorbells in the US, as critics warn it turns every doorstep into a biometric checkpoint.
📌 Key Takeaways
- New Familiar Faces feature lets Ring owners enroll and label up to 50 people for named alerts.
- The tool is opt in, processed in the cloud, and unavailable in Illinois, Texas and Portland due to privacy laws.
- Amazon says biometric data is encrypted, not shared, and unnamed faces are deleted after 30 days.
- Civil liberties groups say it still scans everyone who walks past, risking biometric privacy violations.
- A US senator and digital rights groups are already urging regulators to investigate or block the rollout.
Ring’s Familiar Faces Puts Names To People At Your Door
The new Familiar Faces feature lets Ring video doorbells and cameras do more than tell you that “a person” is outside. Once you opt in, you can build a private gallery of up to 50 faces inside the Ring app.
You tag regular visitors like family, neighbours, or delivery drivers. When any of them appears on camera, Ring can send a notification that says things like “Mum at Front Door” instead of a generic motion alert, and show their name in the event timeline.
The feature is off by default. Owners have to enable it in settings, then manage faces through a new Familiar Faces library, with options to edit labels, merge duplicates, or delete entries entirely.
Amazon says face data is encrypted, processed in the cloud, and that any unlabeled faces are removed after 30 days. Labeled profiles are kept until the owner deletes them. The company also insists the biometric data is not used to train AI models.
Opt In Controls Do Not Calm Privacy Fears
Digital rights groups argue the core problem is not the settings menu, but the way the system works. To recognise anyone, Ring has to scan every face that passes in front of the camera, including neighbours, postal workers, canvassers, and kids who never agreed to be in a private face database.
- Cameras must scan all faces, not just the ones you tag
- Biometric data for non users can be stored for months, critics say
- Laws in some states demand explicit consent for faceprints
- Regulators can hold both Amazon and device owners liable
Earlier explanations of the technology noted that even untagged biometric data could be kept for up to six months, something critics say clashes with state level biometric laws that require explicit, affirmative consent before collecting faceprints.
The feature is also blocked entirely in Illinois, Texas and Portland because of those local rules. That creates a patchwork where a walk past one front door is just a camera event, but a walk past another could generate a biometric record without the person’s knowledge.
“Knocking on a door, or even just walking in front of it, should not require abandoning your privacy.” — F. Mario Trujillo, Staff Attorney, Electronic Frontier Foundation
A US senator has already called on Amazon to abandon the feature, and campaigners are urging state regulators to test their biometric laws against it in court rather than wait for quiet normalisation of neighbourhood face scanning.
Amazon’s Safety Pitch Versus Surveillance Reality
Amazon frames Familiar Faces as a quality-of-life upgrade. The company says it reduces noise by filtering out routine visits and adding context to alerts, so people can review “important moments involving specific familiar people” more easily.
“Familiar Faces empowers customers to reduce notifications triggered by familiar people’s routine activities.” — Amazon
The launch follows a wider AI push across Ring devices, including a Search Party feature that uses neighbourhood cameras to help locate lost pets, and new 2K and 4K models with higher resolution and upgraded on-device processing.
Privacy advocates say combining those capabilities with facial recognition, plus Ring’s history of working with law enforcement, is what makes this moment different. They warn that dense networks of doorbell cameras, each silently building face libraries, can start to look like a distributed surveillance system rather than a personal security tool.
Some early customers are echoing that concern. Reports already describe users cancelling subscriptions over fears that data could be misused, breached, or repurposed in ways today’s policies do not clearly prohibit.
How Familiar Faces Fits Into The Home AI Arms Race
Ring is not the first smart camera brand to offer facial recognition. Rival systems have long supported “familiar face” features, sometimes with on-device storage and stronger limits on cloud use. Amazon is now bringing a similar idea into an ecosystem that already sparked debate over neighbourhood sharing and police access.
For the AI industry, the move shows how quickly agent-like perception is moving into everyday hardware. Devices at the edge are no longer just streaming pixels; they are turning real people into structured biometric data that can be searched, labelled, and linked over time.
Whether this becomes the default for consumer security or a high watermark that regulators roll back will depend on two things: how aggressively watchdogs enforce biometric rules, and how many consumers decide that named alerts are not worth the feeling of being scanned every time they walk past a front door.
Conclusion
Ring’s Familiar Faces feature is pitched as a small convenience upgrade, turning vague motion pings into specific names. Underneath, it quietly shifts millions of doorbells from simple cameras into biometric sensors that process the faces of anyone who comes near.
As AI creeps further into consumer security, this launch is likely to become a test case for how far companies can go in normalising face recognition in everyday spaces.
The next moves from regulators, and from customers who choose to enable or ignore the feature, will decide whether Familiar Faces feels like safety, or like a line crossed.
For the recent AI News, visit our site.
If you liked this article, be sure to follow us on X/Twitter and also LinkedIn for more exclusive content.
