Meta hit with lawsuit over smart glasses facial recognition tech using social media photos

Meta Faces Class Action Over Unreleased Facial Recognition Technology

Meta is confronting a proposed class action lawsuit filed in the U.S. District Court for the Northern District of Illinois over allegations that the company illegally harvested millions of photographs from Facebook and Instagram to develop NameTag, an unreleased facial recognition system designed for its AI-powered smart glasses5

5

. Filed on September 4, 2026, by families in Illinois and California, the lawsuit accuses Meta of extracting biometric data from user-uploaded images without obtaining proper user consent or providing adequate notice, violating Illinois' Biometric Information Privacy Act and California privacy laws1

1

5

5

.

The NameTag feature, discovered in the code of Meta's companion app for its smart glasses, was designed to create biometric signatures—or faceprints—from faces captured by the glasses and match them against a database stored on users' phones1

1

. According to the complaint, Meta had assembled a collection of faceprints on its servers ready for deployment once NameTag was activated, potentially enabling real-time identification of individuals through the smart glasses2

2

. Plaintiffs cite recently filed Meta patents, including one titled "User Identity Verification without Sharing Biometric Data With Platforms," which describes systems where biometric signatures from hardware like smart glasses can be compared against social media data including profile photos, tagged images, and videos1

1

.

AI Training Practices Draw Additional Scrutiny

Beyond the NameTag feature, the lawsuit challenges Meta's broader AI training practices, specifically targeting the company's generative AI models Emu and Muse. The complaint alleges these models were trained on vast quantities of Facebook and Instagram data, including photographs and text, from which biometric information could have been extracted without proper authorization1

1

5

5

. The Muse image model sparked significant backlash when Meta announced it would allow people to generate AI images of others using public Instagram accounts—a feature the company quickly shelved following public outcry2

2

.

This legal challenge raises fundamental questions about data harvesting and how social media images can be repurposed for AI development, particularly when individuals appearing in photographs may not have separately consented to biometric processing5

5

. The allegations suggest Meta's data practices extend far beyond what users might reasonably expect when uploading personal photos to social platforms.

Privacy Violations and Smart Glasses Surveillance Concerns

The privacy violations alleged in the lawsuit extend to a separate but related complaint involving over 70 people who bought, used, or were recorded with Meta's smart glasses4

4

. This 230-page lawsuit, amended in late August, claims the glasses captured intimate and sensitive images—including footage of people undressing, using bathrooms, having sex, and entering passwords—which were then exposed to contractors in Kenya who label content for AI training4

4

. One California user alleged his glasses unknowingly recorded after he placed them on a bathroom counter, resulting in photos of family members bathing appearing in his gallery4

4

.

The lawsuit alleges that every activation of Meta AI—whether intentional or accidental—captures video and audio transmitted to Meta's servers and routed overseas for review by workers who watch, label, and embed these moments into Meta's AI models4

4

. This AI surveillance has contributed to the smart glasses earning the derisive nickname "pervert glasses" as public backlash intensifies3

3

4

4

.

Meta's Response and History of Biometric Settlements

Meta has categorically denied the allegations, stating the lawsuit "is without merit and misrepresents our work."2

2

The company maintains it has been transparent about how user information is used to build and improve AI products3

3

. Regarding NameTag specifically, Meta stated that "nothing has shipped to consumers and no final decision has been made on what to do here, if anything," and denied building a universal face database2

2

3

3

. The company removed NameTag code from the companion app following media reporting2

2

.

However, Meta's track record on biometric data issues complicates its defense. The company previously agreed to a $650 million settlement in 2020 after wrongly collecting faceprints from a billion users and storing them in company databases1

1

3

3

. More recently, Meta paid $1.4 billion in 2024 to settle separate biometric data allegations brought by Texas3

3

5

5

. These settlements demonstrate a pattern that could influence how courts and regulators view the current allegations.

Implications for AI Development and Consumer Protection Laws

The proposed class action could potentially affect millions of users whose images were uploaded to Facebook or Instagram, as well as individuals whose photos were submitted to Meta's generative AI systems5

5

. Plaintiffs are seeking monetary damages and other legal remedies under consumer protection laws from various states4

4

. The case centers not only on Meta's surveillance of people's lives without consent but also on the exploitation of their image and likeness, according to Tina Wolfson, one of the lawyers representing plaintiffs4

4

.

As Meta expands its ambitions in AI and wearable technology, this lawsuit could intensify regulatory scrutiny over how photographs, facial information, and personal data are used to develop emerging AI technologies5

5

. The controversy has already prompted responses from advocacy groups, with the American Civil Liberties Union and more than 70 organizations expressing concerns about the deployment of biometric tools through smart glasses4

4

. The outcome could establish important precedents for how tech companies must handle biometric signatures and obtain consent when developing universal facial recognition systems and training generative AI models on user-generated content.