The funniest thing about smart glasses is that the old idea never really died.
It just waited for AI to make the pitch less embarrassing.
Google is bringing intelligent eyewear back into the main consumer lane. At I/O 2026, the company said audio glasses powered by Android XR and Gemini are coming this fall, with fashion eyewear partners handling the frames and Samsung and Qualcomm involved in the platform stack.
The first version is audio-first, with no big screen in your field of view and no full AR layer trying to turn the world into a dashboard. It is cameras, microphones, speakers, Gemini, and enough app integration to make the phone feel less necessary.
That is the smart entry point.
The assistant gets eyes
Google describes the glasses as a way to get help without pulling out your phone. You can ask Gemini about what you are seeing, get walking directions based on where you are facing, send texts, summarize missed messages, capture photos and videos, translate speech and written text, and hand off tasks like preparing an order before you confirm it.
That is not a new screen.
It is a new sensor surface.
The assistant does not have to wait for you to upload a photo, type context, or describe what is in front of you. It can start from what you are looking at, where you are standing, and which apps are connected.
That makes the product more useful.
It also makes it much more sensitive.
The Google Glass scar tissue is still there
Google knows this category has history.
The original Google Glass failed for plenty of reasons, but one of the big ones was social. People did not like the feeling that a camera was quietly entering normal rooms, normal conversations, and normal public life.
The world is different now because AirPods trained people to talk to invisible assistants, Meta trained the market that camera glasses can look normal, and AI trained users to expect multimodal help. The hardware can be less strange because the behavior around it is less strange.
But less strange is not the same as settled.
When glasses can see, hear, translate, summarize, capture, and route tasks through personal apps, the trust problem moves outside the owner. Everyone nearby becomes part of the product environment whether they bought the device or not.
That is where this category gets hard.
The real signal
Google is not trying to win smart glasses with eyewear alone.
It is trying to make Gemini ambient.
Search puts Gemini in the question box, Android puts it in the phone, Workspace puts it in the inbox and documents, and Android XR puts it on the body.
That is the bigger strategy: fewer moments where the user has to open an app and more surfaces where the assistant is already close enough to act.
The audio-first approach is pragmatic. It avoids the weight, cost, battery, and social weirdness of full display glasses while still putting AI in the physical world.
But the product will live or die on restraint.
If the glasses feel like a helpful earbud with context, people may accept them. If they feel like a camera strapped to every conversation, the old backlash comes back fast.
Gemini on your face is technically possible.
Now Google has to make it socially tolerable.
Sources: Google, TechCrunch