On June 2, 2026, a Virginia man who has never owned a Ring camera sued Amazon in federal court in Seattle, arguing that Ring's Familiar Faces feature scanned and stored his face without asking. The case will take years to resolve, but the argument at its center is simple and, in our view, correct: a doorbell owner can consent to face recognition for themselves, not for the mail carrier, the neighbor's kids or anyone else who walks into frame. This piece explains what Familiar Faces does, why the lawsuit has a real chance, what Ring's history adds, and how to choose cameras that keep video local, including the Matter 1.5 camera standard that finally makes that practical.
What happened: a class action filed in Seattle
According to TechCrunch, the plaintiff, Charles Sigwalt, filed a proposed class action on June 2 in Seattle, Amazon's home jurisdiction. He says he does not own a Ring device and was captured by other people's cameras while visiting friends and family. The complaint claims that "Millions of other Americans passed by a Ring security camera and unknowingly had their facial recognition information collected."
The Register's report lists the claims: violations of Virginia consumer protection, computer crime and privacy laws, plus breaches of Federal Trade Commission standards on "surreptitious and unexpected collection or use of biometric information." ClassAction.org's summary adds that the proposed class covers every U.S. resident whose face data was collected and kept by Familiar Faces, and that the complaint also invokes Section 5 of the FTC Act and state biometric laws such as Illinois's BIPA.
Ring's response, per TechCrunch, is that the feature is opt-in, that face data is "encrypted and never shared," and that "unidentified faces automatically removed after 30 days." A spokesperson told The Register the data is not used to train AI models.
What Familiar Faces actually does, and to whom
Ring announced Familiar Faces on September 30, 2025, alongside its Retinal Vision 2K and 4K cameras and the Search Party pet-finding feature (TechCrunch). It launched in the U.S. in December 2025 and in the UK in April 2026. An owner can build up to 50 profiles of frequent visitors; the camera then labels those people in alerts and, through Alexa+ Greetings, can talk to them by name. Founder Jamie Siminoff's pitch: if someone you do not recognize is lingering, you know immediately.
The mechanism is the problem. To decide whether a face is familiar, the camera must compute a faceprint for every face it sees and compare it against the owner's list. Ring's own support page says unnamed profiles are deleted after 30 days without a recognition, while named profiles stay until you delete them (or, if the subscription lapses, until 180 days pass without a recognition). Those unnamed profiles are the delivery drivers, canvassers and passersby who never agreed to anything.
Opt-in for the owner is not consent from the visitor
This is the legal crux, and the Electronic Frontier Foundation laid it out in November 2025, before launch. In its analysis, the EFF argues that Illinois, Texas and Washington all require affirmative consent before collecting a faceprint, that Colorado and Maryland's newer privacy laws do the same, and that Maryland's effectively prohibits collecting biometrics from bystanders at all (EFF). Its conclusion: "It is impossible for Amazon to obtain consent from everyone—especially people who do not own Ring cameras."
Ring's own behavior supports the point. The feature is not offered in Illinois, Texas or Portland, Oregon, the jurisdictions with the strictest biometric rules (ClassAction.org). The lawsuit argues that carving out those places shows Ring understood the consent problem and shipped anyway everywhere else. Amazon's answer, per the EFF, has been in-app messages reminding customers to comply with local law, which pushes the obligation onto a homeowner who cannot get a signature from the postal worker.
There is precedent that should worry Amazon. The EFF points to Google's $1.375 billion settlement with Texas over cameras that "indiscriminately capture the face geometry of any Texan who happens to come into view, including non-users." The Washington venue matters too: the state strengthened its biometric law in 2023 to allow private lawsuits.
Search Party and the drift from doorbell to neighborhood sensor
Familiar Faces did not arrive alone. Search Party, announced the same day, links outdoor Ring cameras across a neighborhood: when someone reports a lost dog in the Ring app, nearby cameras use AI to look for it and alert their owners, who can then choose to share a clip (Engadget). Dogs came first in November 2025, with other animals to follow.
By February 2026, Biometric Update was reporting that Search Party's ambitions had grown beyond pets, citing a leaked email from Siminoff about using the technology to "zero out crime in neighborhoods." Search Party is on by default. Combine an opt-out neighborhood camera network with a per-camera face database and you have the parts of a distributed surveillance system that no individual customer chose to build. That is the drift the plaintiffs, the EFF and Senator Ed Markey, who opposed the rollout, are pointing at.
Ring's track record: the 2023 FTC settlement and the Flock reversal
Ring's assurances about encryption deserve to be read against its history. In May 2023 the FTC announced a $5.8 million settlement over findings that Ring employees and hundreds of contractors could view, download and share customer video at will, and that more than 55,000 U.S. customers had accounts compromised between January 2019 and March 2020 (FTC). The order mandates a data security program with outside assessments for 20 years.
More recently, Ring canceled a partnership with Flock Safety, the license-plate-camera company that supplies footage to ICE and other federal agencies, amid a backlash that also took in its Super Bowl ad for Search Party (TechCrunch). Ring has repeatedly walked back features after public pressure. That is better than not doing so, but it is not a design philosophy you can rely on.
Our view: convenience features should not require a face database
The honest case for Familiar Faces is that it reduces notification noise and makes greetings feel personal. Both are real benefits. Neither requires building faceprints of strangers.
A camera can tell you "someone is at the door" versus "a package was delivered" using object detection that never identifies a person. It can suppress alerts for a household member by pairing with their phone's presence. Where a vendor does offer recognition, the questions are the same: where is the faceprint computed, where is it stored, and whose faces end up in the database. When the answer is "on a vendor's servers, for everyone who walks past, retained for 30 days for strangers and indefinitely for named profiles," the convenience is not worth the cost, and the cost is paid by people who get nothing in return.
The alternative: Matter 1.5 cameras and local storage
Until recently, choosing a camera meant choosing a cloud. That changed on November 20, 2025, when the Connectivity Standards Alliance published Matter 1.5 with a camera device type (CSA). The spec covers live video and two-way audio over WebRTC, local and remote access via STUN and TURN, privacy and detection zones, pan-tilt-zoom, and "continuous or event-based recording to local or cloud destinations." A Matter camera talks to whichever controller you choose, on your own network, with no proprietary API in the middle.
Samsung SmartThings became the first platform to support Matter cameras in December 2025, with Aqara, Eve and XThings among the partners building them and the first products expected from March 2026 (Samsung Newsroom). For a primer on the standard, see our explainer on what Matter is and why it matters; if you want the controller in your own house, our Home Assistant Green setup guide walks through it.
The practical shift: with a Matter 1.5 camera and a local controller, any recognition that happens can run on hardware you own, with a database you can delete, and no one else's cameras are involved.
A buyer's checklist for privacy-respecting cameras
When comparing cameras, as of mid-2026, ask these questions in this order:
- Can it record locally without a subscription? Look for a microSD slot, a base station with storage, or NAS/RTSP support. If the answer is "cloud only," stop here.
- Does it work without an internet connection? Local recording and live view on your own Wi-Fi should not require the vendor's servers.
- Is it Matter 1.5 certified, or is certification announced? The best guarantee that you can change controllers later without replacing the camera.
- Where is any AI processing done? On-device person and package detection is fine. Face recognition in the vendor cloud is the thing this article is about.
- Does it join any neighborhood or law-enforcement sharing network by default? If so, is opting out one toggle or a maze?
- What is the retention policy for people who are not you? If the vendor cannot answer, that is your answer.
- Is the vendor under a consent order? Ring is, for 20 years from 2023. Not disqualifying, but context.
If you own a Ring: settings to review today
You do not have to throw the doorbell away. Ring's own instructions cover the main controls:
- Delete existing profiles first. Open each Familiar Faces profile and tap the Delete icon; Ring states this permanently removes the profile, the facial recognition information and the images used for it. Do this before turning the feature off, so you can see what was stored.
- Turn off Familiar Faces. In the Ring app, open the menu, go to Pro Features > Familiar Faces and toggle it off for individual devices or all of them. Ring says this stops new faces from being identified.
- Opt out of Search Party. It is on by default, so find its setting in the app and switch it off.
- Review privacy zones. Mask the sidewalk and your neighbor's door so the camera records your property, not the street.
- Check video sharing and law-enforcement request settings in Control Center.
None of these steps help the people scanned by other Ring cameras, which is precisely why the lawsuit exists.
What is still unclear
Whether Sigwalt's Virginia-law claims survive a motion to dismiss, whether a Washington court certifies a nationwide class, and whether Amazon changes the feature before a ruling are all open. Ring's retention figures have also shifted: the EFF wrote in November 2025 that unmatched faceprints were kept for up to six months, while Ring's June 2026 statement and current support page say 30 days for unnamed profiles. That may reflect a policy change after launch; Ring has not said.
Familiar Faces asks a homeowner to make a privacy decision on behalf of everyone who approaches their door, and the lawsuit filed on June 2 argues, persuasively, that the law does not allow that. Whatever the courts decide, you do not need to wait: turn the feature off and delete its profiles if you own a Ring, and for your next camera, choose local storage and a Matter 1.5 device that keeps recognition, if you want it at all, on hardware you control.
Sources and further reading
- Amazon faces class action lawsuit over Ring facial-recognition feature (TechCrunch)techcrunch.com
- Ring faces class action over facial-recognition feature (The Register)theregister.com
- Ring lawsuit claims Familiar Faces feature violates basic notions of consumer privacy (ClassAction.org)classaction.org
- The Legal Case Against Ring's Face Recognition Feature (EFF)eff.org
- Ring cameras can now recognize faces and help find lost pets (TechCrunch)techcrunch.com
- Ring announces Retinal 4K Vision and Search Party (Engadget)engadget.com
- Ring expands Search Party from lost dogs to broader surveillance (Biometric Update)biometricupdate.com
- FTC Says Ring Employees Illegally Surveilled Customers (FTC)ftc.gov
- Familiar Faces (Ring Support)ring.com
- Matter 1.5 Introduces Cameras (CSA)csa-iot.org
- SmartThings first to support Matter cameras (Samsung Newsroom)news.samsung.com