You Can Get Google’s New Smart Glasses Next Month. I’ve Tried Them Out

In Gentle Monster’s hallucinatory showroom, I stared up at four robotic heads turning and talking inside floating spheres. I listened to them speak different languages. All of them were translated in my ears, on the fly, as I wore a pair of Gentle Monster smart glasses, “Intelligent Eyewear” codeveloped by Google and Samsung, which will finally arrive Dec. 10. They’ll be Google’s first pair of smart glasses since its Google Glass days more than a decade ago, and they’re landing at a time when public opinions on AI, smart glasses and data privacy couldn’t be more heated.

Google’s return to smart glasses has been promised for nearly two years, but the first waves of actual prescription-ready eyewear are only now rolling in nearly at year’s end, an unusual time for a product launch. Really, it feels more like a kickoff for efforts that will continue next year, as Google also works on AI-enabled glasses that have displays built in.

I’d worn Google and Samsung’s smart glasses in multiple demos before, but this month for the first time I got to try on and test the final versions designed by Warby Parker and Gentle Monster. I also visited two immersive showrooms the companies are opening in New York City to demo how these glasses work. These showrooms will be open by appointment.

More from CNET

The showrooms were like playgrounds, full of installation art to try the glasses in. Warby’s space was a series of library-themed rooms, with travel themes and whimsical sculptures. Gentle Monster’s was a stark, surreal series of rooms, blending industrial design and realistic (sometimes robotic) horses, people, chickens, floating heads, and … a big blue gentle monster.

The glasses I wore give me a strong sense of deja vu after years of reviewing Meta smart glasses and lots of others, but there are also signs of how their deeper-connected Gemini smarts could be a big difference-maker — and likely a sign of how Apple might approach its own glasses expected in 2027.

Clean, crisp, familiar

The various models of smart glasses across Warby Parker and Gentle Monster designs, which start at $395 and $449 respectively, all work the same: The frames do not have displays, just standard glasses lenses that support a range of prescriptions and lens types that match what Warby and Gentle Monster already offer. They have speakers on the arms that project audio, along with four beamforming microphones that filter out background noise and focus on the direction you’re facing. They have a single camera on the left side, and an LED light indicator on the right.

I got to try on all the frame designs and then wore frames with my prescription lenses popped in so that I could actually use the glasses around the exhibits. The glasses feel reminiscent of what Meta’s made for years, with similar-looking frame size and arms compared to recent Ray-Ban Gen 3 models. Most of the frames weigh around 40 grams, which felt pretty easy on my face, but I wear chunky glasses to begin with.

And the frame designs, while mostly clean and somewhat conservative, all looked really good. Warby Parker’s collection of designs came in several understated color shades (I liked the dark green Pine the best), and familiar round and square frames. 

Cameras come standard

All of the Intelligent Eyewear models across both brands have cameras, a feature that Google’s head of XR, Shahram Izadi, still sees as key to having a deep assistive experience, although he’s also open to the idea of future versions that could be camera-free like Meta’s recent Ray-Ban Audio model.

Cameras have been a big flashpoint in smart glasses over the past year, with concerns over harassment and privacy, certain countries and businesses banning them and worries about them being socially inappropriate.

These glasses use the same type of indicator system as Meta’s glasses do, flashing an LED that also auto-adjusts brightness outdoors and playing a sound when taking photos and videos.

Autorecognizing language translation is wild

I was particularly impressed by how Gemini can just listen to any snippet of language and translate it to what language you need, a trick that makes these glasses feel like a Swiss Army knife for travel. Several demos showed this off, including the surreal room of robotic heads that discussed Gentle Monster’s design philosophy in a handful of languages. I heard them all, sometimes with a short delay or hiccup, over 10 minutes or so.

The glasses can recognize written languages, too, in a much larger range than I’m used to: over 70. Live translation kicks in by pressing and holding two fingers on the glasses temple, and the translated voices change to try to match different speakers in a room as well as varied intonation.

Gemini stays on 

Pressing and holding the side temple turns on Gemini, and once it’s on, it’s on until a significant pause of about 15 seconds or so. That meant that I could keep chatting and asking questions without asking “Hey Google” (or “Hey Gemini,” a new trigger phrase). I could end a session by swiping backward on the arm.

Any of the chats I have can also be turned into a note or summary instantly, hooking into Google Keep. The glasses connect with Google Maps, and Izadi promises that most of Gemini’s key features on phones will carry over to the glasses. On iPhones, the glasses will be able to triangulate across Google apps via the cloud to access Gmail, calendars and more.

I found a few clever demos that showed me some promise. In a pantry installation full of world foods in Gentle Monster’s showroom, I asked Gemini to compare brands of international mayo in front of me. I could point to things and casually ask, “What is this?” And Gemini knew what I was referring to. In another demo, Gemini knew the sheet music I was holding up and played back the matching song. In another demo, I sat down and gave myself a tarot card reading — Gemini could read the cards, regardless of whether the cards were right-side up or reversed.

Other times, Gemini said text I asked it to read was too far away, even though I could read it on a wall. Sometimes it seemed to hallucinate an odd answer to a question. (It responded with “8 minutes” to a question on a wall about whether it was OK to wait 45 minutes for a restaurant table.)

Music playback, photos and videos, too

Tapping the temple arm starts and stops music, swiping changes tracks and two-finger swipes change volume (which was a bit confusing). Meanwhile, a top button on the arm takes photos and videos just like Meta glasses do, but in some demo sessions, it seemed like photos got beamed over to a Pixel phone much faster than they do to Meta’s app.

The music playback sounded good based on some quick tries in a crowded room. The audio seemed particularly loud, maybe even a bit more so than Meta’s glasses, but I’ll know more on that when I review.

How good and reliable will Google’s AI be?

The real test of Google’s Intelligent Eyewear is in how truly well connected those AI services feel and on the battery life. Meta’s up-to-eight-hour battery life on the Scriber Optics glasses I’ve been wearing the last few months tend to drain battery fast in heavy use. Will Google manage battery better, or similarly? 

It’s the AI features and Google app hook-ins that could give Google a leg up. My several hours wearing pairs and wandering through two different immersive public installation spaces only showed me hints, and really, I’m most curious about how the glasses can tap into my entire Gemini history, my Gmail, my Google Docs, or my Google Maps history. How smart and connected can they be, and how good (or unsettling) will that feel?

A shift in privacy

For now, it’s Google’s approach to privacy on its services for these AI glasses that has a chance to stand apart. Cameras are standard for this first wave of glasses, but Gemini treats camera feeds in AI mode differently than how photos and videos are taken otherwise, Google’s Juston Payne, senior director of project management for XR, told me. Always-on Gemini modes with camera don’t share any photos or videos for AI training or for targeting ads against, and none of the camera captures in AI modes can be seen or shared by the wearer or Google, either. Faces are autoredacted using AI on your connected phone before being sent to the cloud for AI analysis, Payne says.

That’s encouraging, but Google’s already raising questions of AI privacy with targeted ads in Gemini-powered Google search on phones and computers.

Google’s also not enabling any livestreaming from these glasses, a move that leans towards careful compared to Meta’s livestreaming modes on its glasses. “We’ll always take the conservative path right now just to rebuild trust,” Payne said. 

The glasses also won’t be able to continuously record audio ambiently to create memory summaries like Apple’s new in-beta Siri Recap feature on its recent watches can or like other smart glasses sometimes do. Also, they don’t have any environmental listening features yet, something Google says is also privacy-focused, but music recognition and other audio awareness features are on the road map. In Gemini-based audio chats with AI, Google also scrubs out voices that might be chatting near you.

But it remains to be seen how much any of this will help make anyone in public feel comfortable with someone wearing a pair of these glasses with camera-enabled AI modes on. It’s already hard to convince people that smart glasses aren’t recording people all the time, and Google’s clearly trying to restore confidence in the idea of smart glasses in public situations.

More glasses on deck

The Warby Parker and Gentle Monster models are not the only new smart wearables Google has coming: A pair of tethered mixed reality glasses that come with a processor puck and work almost like a portable VR headset, called Xreal Aura, is arriving in the next few weeks. And Google’s also starting to test its display-enabled Intelligent Eyewear ahead of a possible launch next year.

Izadi also told me that more assistive features are coming to these camera-enabled glasses, including what Google calls Guided Vision, something of an extension to the Project Astra tech Google has been working on for years. “You’ll hear more about our plans in the coming months,” Izadi said.

But agentic AI is clearly the road going forward, something both Google and Meta are expanding on for glasses. “We’ll bring more and more agentic and core AI features so that people can depend more on the AI capabilities,” Izadi said. Google’s leaning on these glasses being true phone extensions, as well as watch-connected.

But there could be a lot of cascading hardware pieces in between. Google and Samsung have already promised multiple waves of tethered and wireless smart glasses, ones with displays and ones without, and these glasses are apparently just the start.

I’m curious to see how well Warby Parker and Gentle Monster approach support for the glasses and how easy it’ll be to understand how they work. Dec. 10 isn’t far away.

Editor at Large

I started with CNET reviewing laptops in 2009. Now I explore wearable tech, VR/AR, tablets, gaming and future/emerging trends in our changing world. Other obsessions include magic, immersive theater, puzzles, board games, cooking, improv and the New York Jets. My background includes an MFA in theater which I apply to thinking about immersive experiences of the future. See full bio