Published Date : 8/8/2025Â
When a Customs and Border Protection (CBP) officer was filmed wearing Ray-Ban Meta smart glasses during a Los Angeles immigration enforcement action in June, it raised significant concerns about the integration of consumer AI eyewear into federal policing and the Department of Homeland Security (DHS) biometric ID pipelines. The incident not only highlights the potential misuse of advanced technology but also underscores the need for strict adherence to existing regulations.
The officer’s use of the glasses appears to have violated CBP regulations, which prohibit the use of both approved and personal recording devices for capturing individuals engaged in First Amendment-protected activities without reasonable suspicion of an impending enforcement action. The person who recorded the CBP officer does not appear to have been a target of the enforcement action, further complicating the situation.
Neither CBP nor DHS officials have responded to media requests for comments. However, anonymous DHS sources told Biometric Update that there is no official policy specifically for the use of Ray-Ban Meta smart glasses, and existing regulations do not allow for the personal use of recording devices while on duty. One source suggested that the officer likely faced disciplinary action, not for the violation itself, but for getting caught.
This incident is part of a broader trend of increasing surveillance technology adoption by CBP under the Trump administration. This includes facial recognition, access to biometric databases, AI-enabled detection systems, and even technologies like small drones or sensors capable of “seeing through walls.” CBP’s operative policy in this matter is Directive 4320-030B, which was updated in May and has been in effect since August 2021. This directive defines “Incident-Driven Video Recording Systems” (IDVRS) as CBP-owned cameras that are vehicle-mounted, non-integrated vessel systems, and body-worn units.
The policy explicitly states that cellphones are not included in the definition of IDVRS and should not be used as a primary means to record enforcement actions. On-duty personnel are required to use only CBP-issued and approved IDVRS, and no personally owned devices may be used to record law enforcement encounters. Activation of IDVRS should occur at the start of an event or as soon as safely possible, and the system should be deactivated once the officer’s participation concludes.
Supervisors can require a statement if a camera wasn’t activated, and personnel are encouraged to inform those they encounter that they are being recorded when it won’t interfere with safety. The policy also prohibits the use of CBP cameras to film coworkers outside an enforcement encounter, to capture employee assessments outside training, to record privileged conversations, or to film in places with a reasonable expectation of privacy, such as locker rooms or restrooms.
Retention and storage of data are tightly controlled. Recorded data must be uploaded to a designated CBP-approved system and not downloaded or recorded for personal use or posted on personally owned devices or websites. Non-evidentiary data is retained for up to 90 days, while potentially evidentiary material is retained for up to three years, pending National Archives and Records Administration (NARA) approval. Evidentiary data tied to a case file is preserved according to that case file’s records schedule, which can extend for decades.
A complementary Privacy Impact Assessment (PIA) describes the technical stack, noting that beginning in 2021, CBP connected thousands of Border Patrol body cameras to a cloud-based digital evidence platform with role-based access controls and audit logs. While internal planning has referenced a shift from an earlier 180-day non-evidentiary target to shorter operational periods, the publicly posted directive reflects the current retention framework of up to 90 days for non-evidentiary data, a proposed three years for potentially evidentiary material, and evidentiary data retained in accordance with the applicable case file’s records schedule.
Meta’s Ray-Ban glasses, which have a camera, microphones, live-streaming, and “Meta AI with vision,” do not ship with built-in facial recognition. However, multiple outlets have shown it is relatively easy to route the glasses’ video to a phone or laptop and run third-party face-matching pipelines in near-real time. This capability exists in the wild, but it is not an approved CBP function.
Practically, Meta glasses could serve as a capture front-end to CBP’s Traveler Verification Service (TVS), the cloud facial-comparison service that powers Simplified Arrival at airports, seaports, and some land environments. Simplified Arrival uses facial biometrics to automate the manual document checks required for admission into the United States. TVS takes images from approved fixed or partner cameras, matches them against pre-staged galleries built from government holdings, and purges the transient match images from the TVS cloud within defined windows.
None of CBP’s TVS documentation contemplates officer-worn consumer eyewear as a capture device, and the agency’s IDVRS rules would block using personal devices as evidence systems even if the technical and security hurdles were solved. To deploy glasses as a TVS capture source, CBP would need to procure and issue compliant hardware, accredit the software, update PIAs/ System of Records Notices, and revise Directive 4320-030B and component Standard Operating Procedures. There is no indication that this has happened.
The same governance posture appears in DHS’s department-wide body-camera policy, which directs components to operate within strict, agency-owned systems, promulgate detailed component policies, and align training, access controls, and retention with federal records law. This framework leans heavily against ad-hoc recording on personal or consumer devices.
So, what explains the CBP officer wearing Ray-Ban Meta glasses in Los Angeles? The most benign possibility is that they were sunglasses worn as eyewear, with the camera inactive. This is still a bad look for an agency that must project respect for privacy at a time of unprecedented deployment into the nation’s interior. A less benign scenario is off-policy recording, which would collide with the spirit and letter of the IDVRS rule and create a retention, disclosure, and chain-of-custody mess.
If recording occurred, it would also create a retention, disclosure, and chain-of-custody mess because IDVRS footage must flow into CBP’s approved evidence platform with tagging, audit, and NARA-aligned deletion, not into a consumer cloud. Meta’s Ray-Ban Meta smart glasses do not include native facial recognition, but the livestream capability and phone link enable third parties to tap video streams and process them with external AI tools.
In late 2024, Harvard students AnhPhu Nguyen and Caine Ardayfio publicly demonstrated how I-XRAY was used to livestream from the glasses (via Instagram) to a computer program. The system ran facial recognition using PimEyes, then pulled personal information (names, addresses, phone numbers) from public sources and delivered that information back through a phone app in seconds. Nguyen and Ardayfio built I-XRAY to highlight how real-time doxxing is feasible using existing consumer tech.
Publicly documented demonstrations and technical discussions have shown that a real-time face-matching pipeline using Meta’s smart glasses can be built with off-the-shelf tools. The process begins with video capture through the glasses’ livestream feed, which is accessed on a connected phone or laptop. From there, individual frames are extracted and run through face-detection software, such as RetinaFace or YOLO, operating either on the device itself or a tethered computer. Once a face is detected, frameworks like ArcFace or InsightFace generate an embedding, which can then be matched against a prebuilt gallery stored locally or sent to a cloud-based search service like PimEyes. If a match is found, the identity can be enriched with open-source personal data that includes phone numbers, home addresses, and social media profiles before the results are sent back to the wearer.
That feedback may appear in an on-screen app or be spoken aloud through the glasses, often completing the entire process in as little as five seconds. The technology exists today; the glasses simply serve as an inconspicuous capture device. Meta tries to mitigate misuse with a built-in LED recording indicator that can’t be disabled, but in bright light or crowded scenarios, it’s easy to miss. Worse, “stealth” sticker accessories have appeared online attempting to mask the LED, though tests show these often fail or trigger disablement warnings.
For a federal law-enforcement component bound by a specific evidence workflow and a biometric program with its own privacy rules, that plug-and-play path is a legal and policy non-starter. Presently, there is no evidence DHS or CBP has bought Meta’s smart glasses or otherwise integrated Meta platforms into any biometric program.Â
Q: What are the CBP regulations regarding the use of recording devices?
A: CBP regulations, outlined in Directive 4320-030B, prohibit the use of personally owned devices to record law enforcement encounters. Officers must use only CBP-issued and approved Incident-Driven Video Recording Systems (IDVRS).
Q: What happened when a CBP officer was filmed wearing Ray-Ban Meta smart glasses?
A: A CBP officer was filmed wearing Ray-Ban Meta smart glasses during an immigration enforcement action in Los Angeles, raising concerns about the misuse of consumer AI glasses and potential privacy violations.
Q: Can Meta smart glasses be used for biometric identification?
A: While Meta smart glasses do not include native facial recognition, they can be used to livestream video to a phone or laptop, where third-party AI tools can process the video to perform facial recognition and doxxing.
Q: What are the privacy concerns with using Meta smart glasses for law enforcement?
A: The use of Meta smart glasses by law enforcement raises privacy concerns, including the potential for off-policy recording, misuse of personal data, and the circumvention of strict data retention and disclosure rules.
Q: What is the Traveler Verification Service (TVS) and how does it relate to CBP's use of technology?
A: The Traveler Verification Service (TVS) is a cloud-based facial-comparison service used by CBP to automate manual document checks at airports, seaports, and some land environments. It uses images from approved cameras and matches them against government databases.Â