The AG Who Won $1.4 Billion From Meta Targets Its Glasses

Closeup of dark sunglasses reflecting light

TL;DR:

The investigation: Texas AG Ken Paxton issued a Civil Investigative Demand (CID) to Meta on May 20, 2026, targeting the Ray-Ban Meta smart glasses. The CID is a pre-lawsuit tool: it demands internal documents, data practices records, and communications. ^[1][2]
Three targets: The investigation covers facial geometry collection through the glasses’ cameras, the planned “Name Tag” facial recognition feature that could identify strangers on sight, and reports that Sama data annotators in Kenya accessed users’ intimate footage. ^[1][3]
Why Paxton matters: This is the same AG who extracted a $1.4 billion settlement from Meta in July 2024 over Facebook’s unauthorized facial recognition. He’s the most successful biometric privacy enforcer in US history. ^[2][4]
The legal basis: Texas’s Capture or Use of Biometric Identifier Act (CUBI) and the Deceptive Trade Practices Act (DTPA). CUBI carries $25,000 per violation in civil penalties. ^[4]
The bigger picture: There is no federal law governing wearable surveillance devices. Paxton’s investigation is the most aggressive state-level action against smart glasses to date, while 75 organizations have demanded Meta cancel the Name Tag feature entirely. ^[5][6]

What Paxton Is Demanding

A Civil Investigative Demand is not a lawsuit. It’s what comes before one.

On May 20, Paxton’s office sent Meta a CID requiring the company to turn over documents about how the Ray-Ban Meta glasses collect, process, and share user data. The demand specifically targets three areas: the glasses’ capture of facial geometry through built-in cameras, Meta’s representations about privacy safeguards, and the role of third-party contractors who handle the resulting footage. ^[1][2]

“I will continue to relentlessly stand up to any company that threatens the privacy and safety of Texans,” Paxton said in his announcement. “Meta’s glasses raise serious concerns.” ^[1]

Meta responded with boilerplate: “Privacy and data protection are core to every product we build at Meta, including Ray-Ban Meta glasses.” The company said it’s ready to “address the Attorney General’s questions.” ^[1]

That’s roughly what they said last time. It cost them $1.4 billion.

What the Glasses Actually Do

The Ray-Ban Meta glasses ($299/$399) are equipped with cameras, speakers, and microphones. They capture video, audio, and, critically, what Texas calls “facial geometry” of people in the wearer’s field of view. ^[1]

Here’s the privacy design Meta touts: a small LED indicator that lights up when the glasses are recording. The idea is that bystanders can see the light and know they’re being filmed.

The problems with that design are stacking up:

The LED is tiny. It’s small enough to be hidden by a piece of tape, a sticker, or the wearer’s hand. ^[1]
It doesn’t always activate. Meta’s own privacy policy notes the glasses have an “always enabled” mode that constantly processes video data for Meta AI. The LED does not activate during this mode. ^[1][2]
It assumes awareness. Even when the LED is visible, most people don’t know what a glowing dot on someone’s Ray-Bans means. There’s no consent mechanism for the people being filmed.

An app called Nearby Glasses was built specifically to detect when someone around you is wearing Meta smart glasses. The fact that it exists tells you everything about how well Meta’s privacy design is working.

“Name Tag”: Every Wearer Becomes a Walking Surveillance Camera

The CID is about what the glasses do now. But Paxton is also targeting what they’re about to do.

According to a New York Times report based on internal Meta documents, the company is developing a feature codenamed “Name Tag” that would bring facial recognition to the glasses. Two versions are under consideration: one that recognizes people already connected to the wearer on Meta platforms, and another that could identify anyone with a public social media profile. ^[3][5]

Read that second option again. Anyone with a public Instagram account could be identified by a stranger on the street, at a protest, outside a medical clinic, in a bar.

The ACLU’s Cody Venzke put it simply: “Your glasses should not know my name. This is an inherently invasive and unethical technology.” ^[6]

Kade Crockford, director of the ACLU of Massachusetts’ Technology for Liberty program, was more direct: “Stalkers and scammers would have a field day with this technology.” ^[6]

In April 2026, 75 organizations, including the ACLU, the Electronic Privacy Information Center, Fight for the Future, and Access Now, sent Meta a letter demanding the company halt the Name Tag feature entirely. Three US senators, Markey, Wyden, and Merkley, sent their own letter demanding answers. ^[5][6]

Meta’s response: “We do not currently offer facial recognition on our smart glasses. If we were to release such a feature, we would take a very thoughtful approach.” ^[5]

An internal memo reportedly noted that the “political tumult” in the US was good timing for the feature’s release. ^[5] Thoughtful indeed.

The Kenya Problem

While Meta was talking about privacy-by-design, contractors in Nairobi were watching users have sex.

In February 2026, employees at Sama (Meta’s data annotation subcontractor) began speaking to Swedish journalists and Kenyan reporter Naipanoi Lepapa about the footage crossing their screens. They reported seeing Ray-Ban Meta users in bathrooms, in bedrooms, changing clothes, and engaged in intimate acts. The footage was not anonymized before review. Faces were visible. ^[7][8]

These workers were hired to label video data to improve Meta’s AI models. They had no mechanism to flag footage they believed was captured without consent, no way to contact the people being filmed, and no authority to refuse the work. ^[7]

Meta’s response was to terminate its contract with Sama, cutting off 1,108 workers. The people who exposed the privacy violation lost their jobs. The company that created the conditions for it issued a statement about how much it values privacy. ^[7][8]

In March 2026, a class-action lawsuit was filed in California’s Northern District Court against both Meta and EssilorLuxottica (Ray-Ban’s parent company). The 39-page complaint argues that Meta’s “designed for privacy” marketing is a sham. Kenya’s Data Protection Commissioner launched its own investigation. The UK’s Information Commissioner’s Office called the reports “concerning.” ^[8][9]

Paxton’s CID adds Texas to the list of jurisdictions now digging into the Sama contractor revelations.

Why This AG Matters More Than Others

State attorneys general investigate tech companies constantly. Most of those investigations produce press releases, maybe a modest settlement, and nothing changes.

Paxton is different. In July 2024, his office secured a $1.4 billion settlement from Meta over Facebook’s facial recognition system, the largest biometric privacy recovery in US history. The case centered on Facebook’s Tag Suggestions feature, which collected facial geometry from photos without informed consent. ^[4]

That settlement came under the same two laws Paxton is wielding now: CUBI and the DTPA. He’s using the same playbook against the same company, over the same category of violation: unauthorized biometric data collection. The difference is the technology has moved from your phone screen to someone else’s face. ^[4]

Under CUBI, each violation carries up to $25,000 in civil penalties. For a wearable device with millions of users capturing biometric data from unknown numbers of bystanders, the potential exposure is enormous. ^[4]

Paxton’s CID signals he’s doing what he did last time: building the evidentiary record for a major enforcement action. Last time, that process led to $1.4 billion. This time, the underlying technology is more invasive, the data practices are worse, and the planned features are more alarming.

No Federal Law. No Federal Plan.

There is no US federal law that specifically governs wearable surveillance devices. No law requires consent from bystanders whose faces are captured by smart glasses. No law restricts what AI features can be bolted onto cameras you wear on your face.

That vacuum is why state AGs are acting. And it’s why the actions are fragmented.

Texas has CUBI. Illinois has BIPA (which only covers private companies, not government use). New York is building a BIPA-style law. Most states have nothing.

At the federal level, 1,200 AI-related bills have been introduced across the country with no coherent framework. The Trump administration is internally split over whether intelligence agencies or the Commerce Department should evaluate AI models. ^[10] Nobody is talking about the glasses on your neighbor’s face.

The result: if you live in Texas, Paxton might protect you. If you live in a state without a biometric privacy law, Meta can collect your facial geometry through someone else’s sunglasses and there’s nothing you can do about it.

What Happens Next

A CID is the opening move. Here’s what the timeline typically looks like:

Document production: Meta must respond to the CID and turn over internal documents about data collection practices, privacy policies, and the Name Tag feature development
Analysis: Paxton’s office reviews the documents. This usually takes months
Negotiation or lawsuit: Either Meta settles (like it did for $1.4 billion last time) or Texas files suit under CUBI and the DTPA

Meanwhile, the class-action lawsuit in California is proceeding. The 75-organization coalition is pressing for Meta to cancel Name Tag. Three senators are still waiting for answers. Kenya and the UK are investigating. And at RSAC 2026, researchers demonstrated live how easy it is to hack the glasses for facial recognition, even without Meta’s help.

Meta is being squeezed from every direction. The AG who already cost them $1.4 billion just opened a new front.

What You Can Do

If you own Ray-Ban Meta glasses: Disable the “always enabled” mode in settings. Be aware that your footage may be reviewed by human contractors, not just AI. Leave the LED indicator visible and unobstructed
If you’re a bystander: You can’t consent to being filmed by someone else’s glasses in most US states. The Nearby Glasses app can alert you when Meta smart glasses are nearby
If you’re in Texas: File complaints with the Texas AG’s Consumer Protection Division if you believe your biometric data has been collected without consent
Support organizations pushing back: The ACLU, EPIC, EFF, and Fight for the Future are all actively campaigning against wearable facial recognition. The 75-organization coalition letter is still accepting organizational sign-ons
Push for your state: If your state doesn’t have a biometric privacy law, contact your state representatives. Texas’s CUBI is the reason Paxton can act. Most states don’t have an equivalent