A decade after Google launched Google Glass to spectacular failure, the company is trying again. And this time, the world seems more receptive to having an AI assistant like Gemini whisper directions, translations, and reminders directly into your ear. I had the chance to try out a prototype version of Google’s new XR glasses at the Google I/O conference, and they represent a significant evolution in design and public acceptance.
\Design and hardware: Almost indistinguishable from normal eyewear
\The prototype, manufactured by Samsung, blends elements from the Warby Parker and Gentle Monster connected glasses that Google announced separately. The most striking aspect is how normal they feel. There is no extra weight or thickness that would betray the presence of advanced technology. From the front, a small circular notch houses the camera, and a discreet button sits on the right frame for snapping photos. The right arm also features a touch-sensitive surface for swiping, tapping, and volume control via integrated speakers. The speakers were a bit faint in the noisy demonstration environment, but that may be improved in the final consumer version.
\Unlike the original Google Glass, which drew ridicule and spawned the term “Glasshole,” these prototypes look and feel like high-end sunglasses. That shift in design is not accidental—Google has learned that for smart glasses to succeed, they must first be good glasses. The public has also become more accustomed to wearable cameras thanks to Meta’s Ray-Ban Stories and the increasingly normalized use of live-streaming accessories. Still, the presence of a camera and the potential for surreptitious recording remain privacy flashpoints.
\Gemini: An AI assistant that sees the world
\The core proposition of these glasses is the integration of Google Gemini. Rather than requiring constant interaction, Gemini operates in the background and only activates when needed—triggered by a long press on the button or, eventually, a wake word (“Hey Google”). Once active, it can perform the usual AI assistant tasks: play music on Spotify, set reminders, answer factual questions (e.g., “What is ratatouille?”). But the unique capability is that Gemini can also see through the camera and interpret the user’s surroundings. For example, looking at a concert poster could prompt Gemini to check ticket availability, or glancing at ingredients could yield a recipe suggestion. All data travels from the glasses to the user’s phone, which then accesses Wi-Fi or cellular networks.
\This approach eliminates the friction of pulling out a phone, unlocking it, opening an app, and typing a query. It’s a subtle convenience that, over time, could make the glasses an indispensable companion for quick information while walking, driving, or exploring. In my brief trial, the response time was impressive—Gemini interpreted an object and delivered an answer within seconds. However, some use cases feel forced, such as snapping a photo and instantly restyling it into anime. That feature exists but seems more like a demonstration of capability than a practical daily tool.
\The heads-up display: Minimal but useful
\The glasses include a small, integrated heads-up display (HUD) in the upper right corner of the wearer’s field of view. In the prototype, it was rudimentary—showing a simple map or a reminder card. Google explained that the mapping feature adapts based on the user’s perspective, but during the demo, the map was more of an overhead view than a turn-by-turn navigation overlay. I would prefer a guided arrow system, especially for navigating airports, college campuses, or museums. The HUD also displays timers, notifications, and basic information, but it is deliberately minimal to avoid distraction.
\Battery life remains an open question. Google did not provide specifics, but in a device that requires cameras, sensors, wireless connectivity, and a display, power management will be critical. The glasses are designed to connect to the phone, which may offload some processing, but the onboard components still need to last through a typical day. The prototype required periodic recharging during the demo, raising concerns about endurance for all-day use.
\Privacy and social acceptance: Better but not resolved
\Google has addressed some of the privacy concerns that doomed the original Glass. The requirement for a deliberate tap or wake word before recording begins adds a layer of user consent. The company also confirmed that it is not building facial recognition into the glasses—a direct answer to a question I asked during the demo. However, the camera remains capable of capturing still images and video, and I did not see a mandatory recording indicator LED (though Google may add one in the final version). The social stigma around smart glasses has softened, but anecdotal reports of people modifying Meta glasses to disable the recording light show that trust is fragile.
\Another polarizing aspect is the use of a conversational AI assistant in public. Holding a dialogue with Gemini to add a reminder or ask a question while in a face-to-face conversation could be off-putting to some. The glasses transmit audio through bone conduction or tiny speakers, which can be heard by people nearby if the volume is high. Google will need to find a balance between convenience and social etiquette.
\Historical context and future potential
\The original Google Glass debuted in 2012 as a techno-utopian vision of always-connected augmented reality. It failed for three main reasons: high price ($1,500), poor battery life, and a design that screamed “I’m wearing a computer.” This new prototype corrects the design flaw but still faces the battery challenge. The change in public attitude cannot be overstated—Meta’s smart glasses have normalized the form factor, and Google is leveraging that shift. The partnership with Samsung for manufacturing ensures economies of scale, and the later consumer versions from Warby Parker and Gentle Monster will offer prescription lenses and more stylish frames.
\Looking ahead, the potential applications extend beyond the demo. Imagine walking through a museum, and the glasses silently annotate artworks based on your gaze. Or navigating a conference center by glancing at booth numbers. Google is also exploring the integration of real-time translation, which could break language barriers in conversations. The glasses are designed to work as an extension of the phone, not a replacement, but over time, they may become the primary computing interface for specific tasks.
\Mark Hachman, who tested the glasses, noted that the experience felt natural and that the AI was responsive but sometimes forced. He also pointed out that the glasses currently lack standalone capabilities—they rely on a phone connection. This dependency might limit adoption among users who do not want to tether another device. However, for those already invested in the Google ecosystem, the convenience of hands-free AI interaction could be compelling. The final product is expected to launch in late 2026 or early 2027, with a price point likely similar to Meta’s smart glasses (around $299).
Source: PCWorld News