Google’s Gemini is getting a new sense—one that matters in the real world, not just in demos. In an announcement made during Google I/O 2026, Google and Volvo revealed that Gemini will be able to access the external cameras in Volvo’s upcoming EX60 SUV. The goal isn’t to turn the car into a fully autonomous robot. Instead, it’s to give drivers a more capable “explain what I’m seeing” assistant—one that can interpret the vehicle’s surroundings and translate confusing visual information into plain-language guidance.
The first use case Google highlighted is also the most immediately relatable: parking signs. Anyone who has tried to decipher a dense cluster of rules—time limits, permit requirements, loading zones, street cleaning schedules, or odd exceptions—knows how quickly signage becomes a guessing game. Even when the text is technically readable, the meaning can be hard to assemble in your head while you’re maneuvering into a tight spot. With camera-enabled Gemini, Google’s pitch is that a driver can ask a question like “What does this sign mean?” and get an explanation grounded in what the car’s cameras are actually capturing at that moment.
That framing is important. This isn’t simply “Gemini can look at an image.” It’s “Gemini can look at the car’s view of the world,” then connect that view to context the driver cares about. In other words, the assistant is being positioned as a bridge between raw perception and human understanding—especially in situations where the environment is dynamic, the stakes are practical (a ticket, a tow, a missed rule), and the driver needs clarity quickly.
Why Volvo’s EX60 matters: Android Automotive as the on-ramp
Volvo’s role in this story is not just branding. The announcement ties the capability to Volvo’s use of Google’s embedded Android Automotive as its vehicle operating system. That matters because it suggests the integration path is designed to be more direct than a patchwork of third-party apps and loosely connected services.
When a vehicle OS is built around a platform that already supports Google’s ecosystem, it becomes easier to route camera feeds, manage permissions, and coordinate the assistant experience with the rest of the car’s systems. In practical terms, it reduces friction: Gemini can be deployed in a way that’s meant to feel native to the vehicle rather than bolted on after the fact.
This is also where the “sight and mobility” language lands. Cars are uniquely suited to camera-based assistance because they’re already instrumented for vision tasks—parking guidance, obstacle detection, lane awareness, and driver-assistance features. The difference here is that Gemini is being used to interpret what those cameras see in a conversational, explanatory way. The car doesn’t just detect; it can explain.
A new kind of driver assistance: from alerts to interpretation
Traditional driver-assistance systems tend to operate in a narrow band: detect something, warn you, maybe intervene. They’re optimized for safety-critical signals and predictable outcomes. But many everyday driving problems aren’t safety-critical in the same way—they’re comprehension-critical.
Parking signage is a perfect example. A sign might not trigger any immediate “danger” alert, but it can still determine whether you’re allowed to park. The driver’s challenge is not whether the car can see the sign; it’s whether the driver can understand it fast enough to make a decision.
Camera-enabled Gemini reframes the assistant’s job. Instead of only telling you what’s nearby, it can help you interpret what the environment is telling you. That shift—from notification to explanation—could be one of the most meaningful changes in consumer automotive AI, because it aligns with how people actually use cars: not as a controlled environment, but as a constant stream of small decisions.
Think about the moments where drivers hesitate:
Is this a loading zone right now, or only during certain hours?
Does the “permit required” rule apply to my vehicle type?
Is there a time window that starts later than I assumed?
Are there exceptions for residents, EVs, or specific days?
Is the sign partially obscured by a pole, a tree, or another vehicle?
In these cases, the driver isn’t asking for a warning. They’re asking for meaning. Gemini’s value proposition is that it can take the visual input and produce a structured explanation that reduces uncertainty.
Google’s choice of parking signs as the first target is also strategic. It’s a constrained domain with clear inputs and outputs. The assistant can focus on reading and interpreting text and symbols, then translating them into a driver-friendly summary. It’s not trying to solve every driving problem at once. It’s starting with a task that’s both common and measurable: did the assistant correctly interpret the sign?
Of course, the real-world challenge is that signage isn’t standardized everywhere. Fonts vary, languages differ, symbols can be ambiguous, and signs can be damaged or partially blocked. That’s precisely why a conversational assistant approach is compelling: if the system can ask clarifying questions or provide confidence cues, it can handle uncertainty better than a rigid “read this text” tool.
How the experience could work in practice
While the announcement focuses on the capability, the user experience is likely to be shaped by how Gemini is prompted and how it responds.
A plausible interaction pattern looks like this:
1) The driver pulls into a spot or slows down near a sign.
2) Gemini accesses the vehicle’s external cameras.
3) The driver asks a question such as “What does this parking sign mean?” or “Can I park here?”
4) Gemini interprets the sign content and provides an explanation—possibly including the relevant time windows, restrictions, and any exceptions.
5) If the sign is unclear, Gemini may request a closer view or suggest the driver reposition slightly.
This kind of workflow would make the assistant feel less like a passive feature and more like a tool the driver can actively use. It also fits the reality that camera views change with movement. A sign that’s unreadable from one angle might become clear from another. Because the assistant is tied to live camera access, it can adapt to the driver’s position rather than relying on a single static photo.
There’s also a subtle but important benefit: the assistant can potentially incorporate surrounding context. Parking rules don’t exist in isolation. A sign might reference a zone boundary, a curb marking, or a nearby landmark. If Gemini can combine what it sees across multiple frames—rather than treating the sign as a standalone image—it can produce explanations that feel more grounded.
That’s where “interpret” becomes more than “translate.” Translation is converting text from one language to another. Interpretation is converting visual information into actionable meaning. In a car, action matters.
The bigger picture: why “camera-enabled assistants” are taking off
Camera-enabled AI assistants have been trending across consumer tech, but cars are a special case. In a phone app, you can take a photo, upload it, and wait. In a car, you need speed, hands-free interaction, and integration with the driving context.
Cars also create a unique advantage: the camera view is already aligned with the driver’s perspective and the vehicle’s motion. That means the assistant can be continuously “aware” of what’s in front of it without requiring the driver to manually capture images. The driver’s job becomes asking questions, not operating tools.
This is part of why Volvo’s decision to use Android Automotive as its vehicle operating system is more than an engineering detail. It’s a platform strategy. It positions the vehicle as a device that can receive updates and new capabilities over time, rather than a closed system that only supports pre-defined features.
If Gemini’s camera access is rolled out starting with the EX60, it also sets expectations for future models. Once drivers experience an assistant that can explain what it sees, they’ll likely want it for more than parking signs—especially in situations where the environment is complex and rules are easy to miss.
What other applications could follow
Google’s announcement suggests other future applications are possible, though parking signs are the first practical target. It’s easy to imagine a progression that mirrors how drivers encounter visual complexity:
1) Road signs beyond parking
Speed limits, school zones, temporary construction signage, detours, and lane restrictions often require quick interpretation. While many of these are already handled by driver-assistance systems, Gemini-style explanation could help drivers understand exceptions and timing.
2) Lane markings and curb rules
Some curb markings are straightforward; others are not. A conversational assistant could explain what a particular curb color or symbol means, especially when combined with local variations.
3) Vehicle-to-environment “what am I looking at?”
Drivers sometimes see unfamiliar objects—odd bollards, unusual barriers, or temporary installations—and wonder what they are for. An assistant could explain likely meanings and whether they affect parking or driving decisions.
4) Accessibility and compliance cues
In some areas, there are rules about accessible parking, loading zones, or designated areas for specific vehicle types. Gemini could help drivers interpret those rules in plain language.
5) Multilingual support
Even in regions where signage is mostly in one language, tourists and residents alike benefit from translation and explanation. A camera-enabled assistant could reduce language barriers without requiring the driver to know the local terminology.
The key point is that these applications share a common theme: they convert visual information into human understanding. That’s the “sight” part, but the “assistant” part is what makes it useful.
The risks and limitations that will shape adoption
Any system that interprets real-world signage has to deal with accuracy, ambiguity, and safety boundaries. Even if Gemini is strong at reading text, real signage can be messy: glare, rain, low light, occlusion, damaged signs, and unusual layouts. Misinterpretation could lead to a driver making the wrong decision—like parking where they shouldn’t.
So the rollout will likely depend on how Google and Volvo handle uncertainty. For example:
– Does Gemini provide a confidence level or highlight what it’s unsure about?
– Does it encourage the driver to verify with additional signs or curb markings?
– Does it avoid giving overly definitive answers