Live Nov 18 · Cohort 3Nov 18, 19:00 MSK — free hands-on AI agents live workshop · Live Cohort 3 starts Nov 23Join Nov 18 Live →Cohort 3 (Nov 23)
ReleaseInsight · On RadarArs Technica→ original

Google opens preorders for Gemini glasses starting at $395 and blocks covert recording

Google is opening preorders for Gemini glasses starting at $395: the camera lets the assistant see what the wearer sees, while commands can be given by voice. For example, you can look at a poster and ask to add the event to your calendar. There are no displays in the lenses, making results harder to check—and the reliability of actions the assistant takes becomes the central question for products like these. A camera indicator and recording restrictions provide privacy safeguards.

AI-processed from Ars Technica; edited by Hamidun News
Google opens preorders for Gemini glasses starting at $395 and blocks covert recording
Source: Ars Technica. Collage: Hamidun News.
Listen to article
Article Fact Sheet & Key Figures
ParameterDetails
Primary sourceArs Technica ↗
Editor-in-Chief
Key metric #1The camera will not turn on if the recording indicator is covered or modified.
Key metric #2Livestreaming through Play Store apps and facial recognition are prohibited.
Key metric #3Photos and videos will reportedly not be used for training; voice interactions are subject to different terms.

Spot a concert poster, ask to add the date to your calendar, and keep walking. That shorter path from seeing something to taking action is exactly what Google is offering with its first Gemini glasses. There is no longer any need to take out your phone and point its camera: the camera is already on the frame. For developers and product teams, this raises a specific question: how do you trust an assistant to take action when there is no screen in front of you to check the result?

According to Ars Technica, preorders open on November 12, with shipments expected on December 10. Warby Parker frames start at $395, while Gentle Monster frames start at $449. These starting prices include nonprescription lenses. Samsung supplies the computing hardware, while Google handles the software. The frames weigh about 41 grams, excluding lenses.

The camera provides context, while voice triggers action

There are no displays in the lenses. Users summon Gemini by voice or by tapping the frame and hear its responses through open-ear speakers. The assistant can provide personalized answers, read out Google Maps directions, and use images from a 12 MP wide-angle camera.

The poster example neatly explains how the product works. The camera sends an image, and Gemini must read the name, date, and time, understand the request, and create an event. Users skip several familiar steps: taking out their phone, opening an app, photographing the notice, and manually transferring the details. Yet every link in this short chain can fail.

Google also suggests asking for recommendations while looking at a bookshelf. For people with visual impairments, Gemini Guided Vision offers conversational audio descriptions of their surroundings. In this scenario, recognition accuracy and the clarity of responses directly affect how useful the device is.

The Ars Technica author notes that, in his tests, Gemini made mistakes both when identifying objects and when adding appointments. This is the author's observation about the assistant's performance, rather than a measured error rate for the new glasses. Still, it points to a clear interface problem: without a screen, users have a harder time spotting an incorrect date or a mixed-up name.

Camera safeguards combine prohibitions with physical restrictions

Google is making privacy a central part of the launch. The indicator must light up whenever the camera is operating. If it is covered or modified to make the light unnoticeable, the camera will refuse to turn on. The glasses also have a physical switch.

The company prohibits Play Store apps from streaming video from the glasses to social networks, including YouTube. Recorded footage can be saved to the gallery and then uploaded manually. Real-time facial recognition is also ruled out. For developers, these restrictions define the available use cases: livestreaming from the camera cannot be treated as a permitted feature.

The terms of use prohibit recording people without consent and recording in sensitive locations. According to Ars Technica's account, Google reserves the ability to remotely disable the glasses if it identifies a violation. Platform rules can therefore affect the availability of the device itself.

Data promises need to be read separately

Google says photos and videos from the glasses will not appear in Gemini history or be used to train models or target advertising. Voice conversations with the assistant do not receive the same guarantees. It would be a mistake to extend the promise of image protection to every interaction with the device.

According to the Ars Technica report, data is encrypted on the glasses and in transit to the cloud, but that encryption does not extend to Google's servers. The company is working to bring Private AI Compute to the glasses, but that is a future capability. It cannot be counted as a safeguard already in place.

The practical takeaway for teams is to test the entire path from recognition to completed action. For a calendar event, useful safeguards include having the assistant repeat the date aloud, ask for confirmation, and offer an easy undo. Before a pilot, teams will also need to check phone compatibility: restrictions involving iPhone, Pixel, and Samsung Galaxy models have been announced for launch. Warby Parker sales are limited to the US, while Gentle Monster sales are limited to the US and South Korea.

Story

Key Questions Answered

What are the key details on The camera provides context, while voice triggers action?

There are no displays in the lenses. Users summon Gemini by voice or by tapping the frame and hear its responses through open-ear speakers. The assistant can provide personalized answers, read out Google Maps directions, and use images from a 12 MP wide-angle camera. The poster example neatly explains how the product works. The camera sends an image, and Gemini must read the name, date, and time, understand the request, and create an event.

What are the key details on Camera safeguards combine prohibitions with physical restrictions?

Google is making privacy a central part of the launch. The indicator must light up whenever the camera is operating. If it is covered or modified to make the light unnoticeable, the camera will refuse to turn on. The glasses also have a physical switch. The company prohibits Play Store apps from streaming video from the glasses to social networks, including YouTube. Recorded footage can be saved to the gallery and then uploaded manually. Real-time facial recognition is also ruled out.

What are the key details on Data promises need to be read separately?

Google says photos and videos from the glasses will not appear in Gemini history or be used to train models or target advertising. Voice conversations with the assistant do not receive the same guarantees. It would be a mistake to extend the promise of image protection to every interaction with the device. According to the Ars Technica report, data is encrypted on the glasses and in transit to the cloud, but that encryption does not extend to Google's servers.

ZK
Zhemal Khamidun — CPO at Alpina Digital, product manager of AlpinaGPT. Building AI products since 2019. Hamidun News is a pet project that grew into a production aggregator with 6,000+ readers.
✦ Unified AI Workspace · Nomad

Test Frontier LLMs & Autonomous Agents in One Workspace

Run Claude Opus 5.5, GPT-6, Gemini 3.1, and DeepSeek side-by-side in Nomad AI (nomadnet.ai) — unified model routing, deep research, and multimodal workflows under one subscription.

Try Nomad AI →Launch Nomad Agent
What do you think?
Loading comments…