Amazon's Ring doorbell can now identify visitors using a new AI-powered facial recognition feature called "Familiar Faces," the company announced on Tuesday. Initially unveiled in early September, the controversial capability is now being rolled out gradually to Ring users across the United States.
According to Amazon, the feature lets users create a catalog of up to 50 frequently seen individuals—such as family members, friends, neighbors, delivery personnel, or housekeepers. Once a person is labeled in the Ring app, the system will recognize them whenever they appear in front of a Ring camera.
Instead of receiving generic alerts like “Someone is at your door,” users will get personalized notifications such as “Mom is at the front door,” the company explained in its announcement.
The feature has already drawn criticism from consumer advocacy groups like the Electronic Frontier Foundation (EFF) and U.S. Senator Ed Markey.
Amazon notes that users can disable alerts for specific individuals—for instance, to avoid notifications when they themselves enter or exit their home. These preferences can be customized on a per-face basis.
Familiar Faces is disabled by default and must be manually enabled through the Ring app settings.
Users can name recognized faces directly from the event history section or via the new Familiar Faces library within the app. Once labeled, these names will appear consistently across notifications, the app timeline, and event history. Labels can be edited at any time, and tools are available to merge duplicates or remove faces entirely.
Amazon asserts that all facial data is encrypted and never shared with third parties. Additionally, unlabeled facial images are automatically deleted after 30 days.
Privacy Concerns Over AI-Powered Facial Recognition
Despite Amazon’s privacy assurances, the introduction of this feature has raised significant concerns.
The company has a documented history of collaborating with law enforcement, including granting police and fire departments the ability to request footage directly from Ring owners via the Neighbors app. More recently, Amazon partnered with Flock Safety, a manufacturer of AI-driven surveillance cameras used by federal agencies and Immigration and Customs Enforcement (ICE).
Ring’s own security practices have also come under scrutiny in the past.
In 2023, Ring was fined $5.8 million by the U.S. Federal Trade Commission after it was revealed that employees and contractors had broad, unrestricted access to customer video feeds for years. The Neighbors app also exposed users’ precise home addresses and locations, while Ring account credentials were found circulating on the dark web for an extended period.
Given Amazon’s track record of cooperating with law enforcement and surveillance vendors—and its history of security lapses—we recommend Ring users exercise caution when labeling individuals with real names. Better yet, consider keeping the feature turned off and simply reviewing footage manually. Not every feature needs an AI upgrade.
Due to privacy concerns, Senator Ed Markey (D-Mass.) has urged Amazon to abandon the feature altogether, echoing calls from organizations like the EFF. Moreover, biometric privacy laws currently prevent Amazon from launching Familiar Faces in Illinois, Texas, and Portland, Oregon, as noted by the EFF.
In response to these concerns, Amazon stated that users’ biometric data is processed in the cloud and emphasized that it does not use this data to train its AI models. The company also claimed that, even if requested by law enforcement, it technically cannot determine all locations where a detected individual has been identified.
However, given the similarity to Ring’s existing “Search Party” feature—which leverages the entire neighborhood network of Ring cameras to locate lost pets—it remains unclear why such location tracking wouldn’t be feasible here as well.
“Knocking on a door—or even just walking past one—shouldn’t require surrendering your privacy,” said F. Mario Trujillo, a staff attorney at the EFF.
