Meta Plans Private AI Processing for Ray-Ban Glasses, With No Release Date

Meta says Private Processing would keep cloud AI requests out of its own reach. It has not said which glasses features would use the system.

By 2 min read
Meta Plans Private AI Processing for Ray-Ban Glasses, With No Release Date
Meta Plans Private AI Processing for Ray-Ban Glasses, With No Release Date

Listen to this story

The audio brief

About 1:25
0:001:25
Read transcript
Meta plans to bring its Private Processing system to Ray-Ban smart glasses, but owners can’t use it yet, and the company has given no release date. The proposal is about cloud AI: Meta says requests handled through the system would be inaccessible to Meta and other outside parties. Meta announced the plan at Meta Connect. Private Processing is already used in WhatsApp and the Meta AI app, and Meta says the underlying platform is working. But that doesn’t mean glasses features are ready, or that the system has passed an outside audit. Meta says it wants security and privacy researchers, along with a third-party auditor, to examine the platform before it puts more experiences on it. The examples Meta has mentioned include translating a conversation and summarizing notes or a conversation. Those are possibilities, not confirmed features. Meta hasn’t said whether assistant chats, transcripts, or messages will use Private Processing either. Until it names the requests covered, glasses owners won’t know which AI interactions get the promised protection. There’s also a separate privacy question. Cloud-data protection doesn’t establish whether people nearby are recorded or have a say in what the glasses capture. WIRED has reported continuing concerns about face recognition and misuse of the glasses. For now, the key constraint is simple: Meta has announced a privacy plan, but not which glasses features it will protect—or when it will arrive.

Story brief

3 key points

Meta is extending its Private Processing design—already used in WhatsApp and the Meta AI app—to Ray-Ban smart glasses, but owners cannot use it yet and Meta gave no launch date. The company says the cloud platform works and wants security researchers and a third-party auditor to examine it before adding experiences; that is a planned review, not a completed audit. No glasses requests are committed to the system....

  1. 01

    Meta announced the plan at Meta Connect on September 23.

  2. 02

    Translation and summaries of notes or conversations are examples, not confirmed features; assistant chats, transcripts, and messages are also uncommitted possibilities.

  3. 03

    WIRED notes ongoing concerns about face recognition and misuse of the glasses, issues distinct from access to cloud-processed data.

Meta wants its Ray-Ban smart glasses to answer AI requests in the cloud without giving Meta access to the user’s data. At Meta Connect on September 23, the company said it plans to bring Private Processing to the glasses. The proposal is new; the protection is not available on them yet.

An app privacy system heads toward glasses

Private Processing was already in WhatsApp and the Meta AI app last year. Meta describes it as a way to run AI work in the cloud while keeping user data inaccessible to Meta and other outside parties. Bringing that design to glasses would extend the same stated protection to a device that can take in information from a wearer’s surroundings.

Meta points to tasks such as translating a conversation or summarizing notes or a conversation as reasons someone might use AI glasses. Those jobs can involve personal details. Its proposal addresses who could access data sent for cloud AI processing; it does not, by itself, answer every concern about what a pair of glasses might capture in the first place.

A working platform, not a feature list

Shah told WIRED that Meta is speaking about the platform now so it can work with security and privacy researchers and a third-party auditor before putting more experiences on it. That makes outside scrutiny part of the stated next step, not evidence that an audit has been completed or that any glasses feature is ready to ship.

Private chats with the glasses’ AI assistant, transcriptions and messages are conceivable uses, but Meta has not committed to any of them. Shah called the possibilities examples rather than concrete features. Until Meta names the requests that will use Private Processing, wearers cannot tell which AI interactions would receive its promised protection.

The question beyond cloud access

The distinction is important for a product worn around other people. WIRED describes continuing concern about Meta’s smart glasses, including face recognition and how the glasses might be misused. A promise that Meta cannot access data processed for an AI request is different from a promise about whether people nearby are recorded or have a say in that request.

For now, Private Processing remains a plan for the glasses, with no release timeline. Meta’s next substantive disclosure would be which features use it and how the promised review shapes their rollout. Until then, the announcement sets out a privacy goal rather than a protection that Ray-Ban Meta owners can use.

Sources

  1. wired.comMeta Pinky Promises Its Smart Glasses Will Be Private Soon

Loading discussion...

YOUR READING SPACE

Notifications