General Motors is taking a big swing at making your car feel less like a hunk of metal and more like a talkative, context-aware companion. At its GM Forward media event, the automaker announced that it will begin rolling out a conversational assistant powered by Google’s Gemini models to vehicles starting in 2026 — not as an isolated feature, but as an over-the-air upgrade that can reach millions of OnStar-equipped cars going back years.

Think of it as Google Assistant on steroids. GM says the Gemini-based system will be able to do the usual infotainment chores — navigation, messages, playing music — but with much more natural language understanding. The company promises the assistant will also be able to tap into vehicle data (with the driver’s permission) to flag maintenance concerns, toggle climate control remotely, explain car features, and answer location-based questions by searching the web. In short: the car will know more about itself and, for a time at least, rely on Google’s language smarts to talk you through things.

“One of the challenges with current voice assistants… is that they’re trained on certain code words or they don’t understand accents very well,” GM VP Dave Richardson told TechCrunch — a concise way of framing what many drivers already complain about: voice assistants that need you to speak like a robot. Gemini, Richardson said, is less brittle, which is precisely the selling point GM is leaning on.

A Play Store upgrade for cars (including older ones)

A surprise in the announcement: GM won’t limit Gemini to brand-new models. The automaker plans to make the Gemini assistant available via the Google Play Store as an over-the-air update for OnStar-equipped vehicles from model-year 2015 onward. That means millions of drivers with otherwise “older” infotainment systems could get access to the new conversational layer without buying a new car. It’s a practical move — and one that expands Google’s and GM’s footprint in vehicles far faster than waiting for new hardware cycles.

While Gemini will replace the existing “Google built-in” experience in many GM vehicles, the company says this is still an interim step: GM is simultaneously building its own vehicle-centric AI that it says will be tailored to cars and drivers and will launch sometime after the Gemini rollout. For now, Gemini is the bridge to more capable, conversational vehicle assistants.

Privacy is the elephant in the back seat

Announcements about helpful, always-listening assistants are usually followed quickly by privacy questions — and GM’s recent regulatory history has made those questions unavoidable. In January 2025, the Federal Trade Commission acted against GM and OnStar over the sharing of precise geolocation and driving behavior data without clear consumer consent, and the company agreed to restrictions on sharing that data for five years. That history looms over any new feature that promises to access vehicle telemetry to deliver personalized help.

GM is trying to head off concerns by saying Gemini’s integration will be “privacy-focused” and will let drivers control what the assistant can access. But promises of granular controls will be judged by implementation: whether settings are understandable, whether defaults favor privacy, and whether drivers — especially owners of older cars getting an OTA update — are properly informed and given genuine choice. The FTC settlement raises the bar for transparency here; regulators and privacy advocates will be watching closely.

The wider tech roadmap: eyes-off driving, new compute, and goodbye CarPlay/Android Auto

Gemini wasn’t the only headline at GM Forward. The company also previewed an “eyes-off” highway driving system slated for 2028 — a step beyond today’s hands-free systems — and announced plans for a new centralized vehicle computing platform in the same timeframe. Those moves make the Gemini rollout feel like one piece of a larger transition from hardware-defined cars to software-first vehicles.

Another practical consequence: GM says it will phase out Apple CarPlay and Android Auto as it moves toward its own unified software strategy. In other words, the days of plugging your phone into the dash to get a familiar interface are numbered on new GM releases — and that change will accelerate as the company pushes more functionality into built-in software and AI assistants. For drivers who prefer third-party ecosystems, that’s a meaningful shift.

Why it matters (and what to watch)

There are two big reasons this matters beyond the showroom speeches. First, making advanced AI available through an OTA update to older vehicles lowers the cost and friction of adoption. It’s one thing to promise futuristic cars in 2030; it’s another to put conversational AI into cars people already own. That could change how drivers interact with their vehicles — for better or worse.

Second, the partnership underscores a broader contest between automakers, Big Tech, and chip/software vendors over who owns the in-car experience. If your car’s assistant is provided and updated by Google (and eventually by GM itself), the economics of apps, data, and services shift away from phone ecosystems and toward vehicle makers and platform partners. That raises questions about competition, data governance, and what choices consumers will retain.

Starting in 2026, some GM drivers will get a significantly smarter voice assistant — one that can reach onto the web, access vehicle diagnostics, and (if you allow it) act on things like climate control from outside the car. It’s a major step in the slow remaking of cars as software platforms, but it comes at a moment when trust in how automakers handle driver data is fragile. If GM wants drivers to hand over control — even a little — the company will have to show it’s learned from past mistakes, make privacy controls obvious and meaningful, and ensure the assistant actually makes life easier rather than just louder.