California Just Started Ticketing Driverless Cars — And It's a Bigger Deal Than It Sounds

For years, when a robotaxi blew through a red light or parked itself in front of a fire truck, the answer to “who gets the ticket?” was painfully simple: nobody. No driver, no citation. California just closed that loophole — and in doing so, started writing the playbook for how the law treats AI in the physical world.

The Bay Area Has Been Waiting for This

If you’ve driven in San Francisco in the last two years, you’ve probably seen it. A Waymo frozen mid-intersection. A Cruise vehicle blocking an ambulance. Crowds filming as a driverless car ignored a cop’s hand signals. Police could do exactly one thing about it: call the operator’s customer service line.

That’s not a hypothetical frustration. A YouTube video from attorney Steve Lehto unpacking the new rule racked up over 27,000 views in a single day — a strong signal that this nerdy DMV update touched a nerve. San Francisco residents and first responders have been complaining for years that driverless vehicles operated in a kind of legal twilight zone, where the normal accountability tools simply didn’t apply.

What Actually Changed

The new California DMV regulation does one deceptively simple thing: it lets law enforcement issue traffic citations directly to the company operating the vehicle when a driverless car breaks the rules. Waymo, and any other operator running fully autonomous vehicles on California roads, are now on the hook the same way a fleet operator would be.

The reason this took so long is structural. Traffic law in the US was written assuming a human behind the wheel. Every citation, every point system, every license suspension — all designed around an individual driver. Pull that human out, and the entire enforcement machine grinds to a halt. California didn’t rewrite the whole machine. It just said: the operator is the driver now.

Why Citations Hit Harder Than Lawsuits

You might think a few traffic tickets are nothing to a company valued in the tens of billions. That misses the point.

Citations create a paper trail. Each violation becomes a data point the DMV can use when reviewing operating permits — and California is the agency that grants those permits in the first place. A pattern of running stop signs or impeding emergency vehicles isn’t just embarrassing PR anymore. It’s documented evidence that an operator’s AI system has measurable failure modes, sitting in a regulator’s file.

That changes the incentive structure. Lawsuits are slow, settle quietly, and disappear under NDAs. Citations are public, cumulative, and tied directly to the license that lets you operate at all. For the first time, “our software made a mistake” has a price tag the regulator can ratchet up.

The Template for Everything Else

This is where the story gets bigger than driverless cars.

We are about to spend the next decade arguing about who is responsible when AI agents execute trades, when robots damage property, when an algorithm denies someone medical care. The legal frameworks for all of that are missing or patchwork. California’s move matters because it picked the most visible, most legible domain — traffic law — and established a clean principle: the entity operating the autonomous system is the legally responsible party.

Other states will copy this. They almost always copy California’s vehicle rules, because California is the market everyone has to ship into. And once that principle is normalized for cars, applying it to delivery robots, autonomous drones, and eventually to non-physical AI agents becomes a much shorter argument.

The Takeaway

There is no driver behind the wheel, but there is still a driver. That’s the legal fiction California just made real. It’s a small thing — a few traffic tickets in San Francisco — and it’s also the start of a much longer conversation about what accountability looks like when the thing breaking the rules isn’t a person at all. The question isn’t whether the rest of AI regulation will follow this shape. It’s how quickly.