Loading story…
Meta has announced plans to extend its Private Processing architecture to AI glasses, building on the approach it previously outlined for WhatsApp AI tools. The architecture is designed to protect sensitive cloud-based AI workloads and persistent personal context using hardware-backed confidential virtual machines, remote attestation, anonymous credentials, third-party Oblivious HTTP (OHTTP) relays, and encrypted storage. Meta says these safeguards are intended to prevent it from accessing protected user data. Before transmitting personal context, the glasses verify the cloud workload against an append-only transparency ledger and establish an attested TLS session; data is withheld if verification fails.
Deployment remains uncertain: Meta says the platform is ready and working, but WIRED reported that Private Processing is not yet available on the glasses. Meta has not announced an availability timeline or identified the first supported features. The company is engaging independent security firms, including NCC Group, researchers, and a third-party auditor to review the architecture. For enterprise assessments, the distinction between announced protections and deployed capabilities remains important, while persistent wearable sensing continues to raise privacy concerns that cloud-processing safeguards alone do not resolve.
Track how attackers are adapting to this technology.
3 events from the most recent confirmed update back to the earliest known activity.
Meta introduced its Private Processing architecture for WhatsApp and Meta AI to protect cloud-based AI workloads.
Meta said independent security firms, including NCC Group, and security researchers were reviewing the architecture and attestation mechanisms. Engineering director Pritam Shah said Meta announced the working platform before broader deployment to engage security and privacy researchers and a third-party auditor.
Meta announced plans to extend Private Processing to AI glasses for sensitive cloud-based workloads and persistent personal context. The disclosed architecture combines confidential virtual machines, remote attestation, a transparency ledger, anonymous credentials, third-party OHTTP relays, and encrypted storage; Meta did not announce an availability timeline.
Vulnerabilities, threat actors, malware, products, organizations, and breaches Mallory has linked to this story.
Follow how adversaries are adapting to this technology, and where it touches your stack today.
2 references tracked. Mallory keeps watching after this page renders.
Map indicators from this story to your assets and identify affected systems in minutes.
Every observed campaign, victim, and pivot linked to actors named in this story.
Malware, exploits, and IOCs connected to the activity described here.
YARA, Sigma, and Snort rules deployed to your SIEM as soon as they’re published.
Get matching new stories delivered to your team as they break — not the next morning.
Ask questions about this story and take action on the answers.
Other stories Mallory is tracking that overlap on entities, sources, or topic labels.
Leaked documents and app analysis indicate Meta was developing a facial-recognition feature called **NameTag** for its Ray-Ban smart glasses and the Meta AI companion app. Reporting cited in the references says the software included on-device components for face detection, face cropping, liveness checks, and biometric template generation, with the apparent goal of identifying people seen through the glasses by linking them to Meta platform data such as Facebook and Instagram profiles. The code was reportedly distributed inside the Meta AI app to millions of smartphones even though the feature was not publicly launched. After scrutiny and backlash, Meta reportedly removed most of the related code from a newer version of the app, but privacy concerns persisted because the capability could be restored and because bystanders would have no meaningful way to opt out of being scanned. The reports also say Meta licensed facial-recognition and liveness-detection technology from **Rank One Computing**, a vendor known for government and law-enforcement work, and that the software could support databases of up to **10 million** biometric profiles, intensifying concerns about consumer surveillance through wearable devices.
Jul 15, 2026
Meta's Ray-Ban smart glasses are facing growing restrictions from pubs, restaurants, and theatres as venue operators respond to privacy concerns over devices that can discreetly record audio and video while resembling ordinary eyewear. UK organizations including **Soho House**, **ATG Theatres**, and **Wetherspoons** have reportedly banned or limited the glasses, citing risks to patrons and staff, while reports also said footage captured through Meta's AI features was sent to Meta servers and reviewed by human contractors at **Sama** in Nairobi for AI training, including highly sensitive scenes such as people undressing and visible bank cards. The backlash has intensified after a newly published Meta patent application described smart glasses that could pair AI with facial recognition to identify nearby people and automatically generate media content about them. Although the filing does not confirm the feature will be released, privacy advocates warned it could enable stalking, harassment, and abuse, especially against vulnerable people, and renewed scrutiny of Meta's earlier work on facial-identification features such as the internally developed **"Name Tag"** capability and previously disabled facial-recognition code in Meta AI.
Aug 18, 2026
Investigations reported by Swedish outlets *Svenska Dagbladet* and *Göteborgs-Posten* found that recordings captured by **Meta Ray-Ban smart glasses**—including video and audio—are being reviewed by human contractors as part of AI training and quality assurance workflows. Workers employed by **Sama**, a Meta subcontractor in **Nairobi, Kenya**, described routinely handling highly sensitive content inadvertently recorded by users, including bathroom visits, undressing, sex/pornography, and private conversations, as well as incidental capture of **bank cards** and other identifying details; interviewees said they feared reprisals for raising concerns and described strict on-site controls intended to prevent leaks. Following the reporting, the UK’s privacy regulator, the **Information Commissioner’s Office (ICO)**, confirmed it is contacting Meta to ask questions about the devices and associated data-handling practices. While Meta’s terms reportedly disclose that some interactions may be reviewed by humans to improve the system, the reporting and worker accounts suggest the review pipeline can include intimate or identifying moments that wearers may not expect to be viewed by third parties, raising regulatory and reputational risk around consent, transparency, and safeguards for bystander and user privacy.
Mar 5, 2026