Austin has had quite the experience.
Federal safety regulators are demanding answers from Waymo, Tesla, Zoox and other autonomous-vehicle developers after documenting a pattern of driverless vehicles interfering with first responders.
Last week, the National Highway Traffic Safety Administration ordered the companies to explain by the end of July how they will keep their vehicles from entering emergency scenes, blocking responders or failing to recognize flashing lights and other hazards. The incidents have renewed questions about whether robotaxis are safe enough to operate on public streets and prompted calls from some experts to slow their expansion.
Austin has seen such incidents firsthand. In one high-profile case, first responders heading to a March 1 mass shooting downtown faced an unexpected obstacle: self-driving robotaxis.
The shooting left three people dead and the gunman was fatally shot by police. Fifteen others were injured.
As first responders converged on the area, as many as five autonomous vehicles froze near the scene. One sat perpendicular across the roadway for nearly a minute before a police officer was able to climb into the front seat and move it.
Austin-Travis County EMS officials have said the mishap did not affect patient outcomes.
But robotaxis either blocked traffic or ignored the direction of Austin police in 25% of reported incidents from July 2023 to June 2026, according to the city of Austin’s autonomous vehicle database. Data collected by the city since 2023 shows interference with law enforcement and other traffic is among the top safety concerns surrounding autonomous vehicles in the city.
Experts say the incidents have helped highlight the limited power Texas cities have to regulate the technology on their own streets.
“If Waymo or another company’s autonomous driving system can’t handle flashing lights and sirens in a crisis, it isn’t ready for public roads,” said Cooper Lohr, a senior policy analyst for transportation and safety at Consumer Reports. “It should be removed from service until the company proves it will handle the situation appropriately.”
[…]
Federal regulators are scrutinizing several self-driving car companies operating in Austin over a number of other issues.
Tesla’s Full Self-Driving technology is under investigation over collisions in low-visibility conditions, including a crash in which a pedestrian was fatally struck. The investigation covers incidents involving consumer vehicles, but Tesla notably uses Full Self-Driving technology in its robotaxi service.
Tesla’s technology relies mainly on cameras, instead of additional technology such as lidar and radar sensors that other companies, including Waymo, Zoox and Avride, rely on. But even with overlapping sensors, other companies are facing investigations over safety concerns.
The National Transportation Safety Board and National Highway Traffic Safety Administration are investigating Waymo after vehicles failed to stop for school buses with extended stop arms on more than 25 occasions since August. In at least one instance, a remote human operator told the car to pass the bus.
The company later issued a nationwide recall after one of its vehicles was swept away in San Antonio floodwaters.
Avride is being probed after 16 crashes in Dallas and Austin in which it failed to detect other vehicles or objects. The Austin-based company is testing its autonomous vehicles in the city but has launched public services in Dallas on the Uber app.
In all, Avride has reported 24 crashes in Austin since June 2025. Tesla’s count stands at 20, and Zoox has had two crashes. Waymo, with it’s much larger fleet, has seen 113 crashes in the same period.
Amazon-owned Zoox, which recently began testing its purpose-built robotaxis in Austin, said it has engaged with first responders and city officials for training, especially because its vehicles do not have a steering wheel or foot pedals. The company plans to begin offering rides to a limited group by the end of the year.
The pitch for self-driving vehicles is to make roads safer, Center for Auto Safety executive director Michael Brooks told the Statesman. But the technology has not yet progressed that point, he said.
“These companies want growth to happen quickly, and they want to scale their operations very, very quickly in order to appeal to shareholders and investors,” Brooks said. “But frankly, it’s dangerous. They want to scale as rapidly as possible. But we need a slower and more safety-focused approach.”
Amid a push for rapid expansion, Austin officials say the companies have been reluctant to engage publicly with local leaders — and under Texas law, they are not required to.
It’s a long story and a gift link, so read the rest. This is in some ways a higher-stakes version of the sleeper problem, which is annoying and wasteful but not dangerous. I’ve blogged about a number of the things discussed in this story, but it’s something else to see it all put together in one place.
It’s not clear what cities can do about a lot of this. As noted, cities have basically no regulatory control over rideshare companies or autonomous vehicles. It’s all on the state, and we know how much they care about stuff like this, not to mention how much funding they have for enforcement. The companies have responded as you’d expect in the absence of any incentive, giving lip service to things like training for law enforcement on how to interact with their robotaxis, and then not following up on it. One Austin City Council member suggested in the story that they could jack up the fines for violations involving robotaxis, to something like $50K per ticket, on the theory that it might at least get their attention. I like the idea, but suspect it wouldn’t survive first contact with judicial review. Still seems worth a shot. Maybe Houston can benefit from Austin’s experience, because all of this is going to ramp up here as well. It’s going to be a bumpy ride.