John Lorinc is a freelance journalist specializing in cities, climate, and technology. He can be reached at lorinc@rogers.com.

On a dry morning this January, a nine-year-old girl hopped out of a vehicle in front of her Santa Monica school and darted across the street. At that moment, a Waymo robotaxi traveling in the other direction detected the child and braked heavily. The car knocked the girl to the ground, but she got up and hustled back to the curb. The Waymo stopped on impact and parked until the police arrived. (A National Transportation Safety Board investigation is pending.)

This collision was one of several headline-generating incidents involving the Alphabet-owned fleet. And although industry data, mainly from Waymo, suggests robotaxis tend to be less prone to distraction, breaking traffic laws, and collisions than old-fashioned human drivers, their growing presence on city streets has raised concerns among city officials. Many have observed them block emergency vehicles, endlessly circle suburban cul-de-sacs, or drive through flooded streets. Others are monitoring traffic tie-ups at curbs in heavily trafficked areas, such as outside theaters when a show is letting out.

“City staff have a lot on their plates already,” says Ramses Madou, division manager of planning, policy, and sustainability for San José’s transportation department. “When new technology enters our sphere, it’s important that we address it to the degree that our organization’s goals and policies call for, without getting distracted from other important work.”

While cities don’t regulate vehicles, they do have a broad mandate to manage the road network, which may include the ways robotaxis use city streets. For example, trucks and other large vehicles cannot drive through some residential neighborhoods, while many cities ban airport limos from picking up fares on return trips.

Alexandre Bayen, a UC Berkeley professor of engineering and computer science, points out that earlier transportation technologies, such as routing and ride-hailing apps, significantly affected road use, frustrating city officials. “These problems have not gone away,” he says. “It’s still an issue. Robotaxis are an additional service, which also uses automated routing, just like with ride-hailing apps. Hence, more of that is going to happen.” 

Other experts warn that steadily increasing use of conventional robotaxi services will result in more vehicle travel. Daniel Sperling, founding director emeritus of the Institute for Transportation Studies at UC Davis, says research has shown that ride-hailing vehicles travel about 40% more than ordinary passenger vehicles because of all the time they spend driving around between fares. “In the cities at least, you’re taking away some of the transit patronage,” he adds. “That’s what the cities worry about. The effects will be magnified with personally owned automated vehicles because then people will use the vehicles as offices and hotel rooms and for entertainment.”

Yet some cities see robotaxis and, eventually, autonomous micro-buses as a promising way to improve transit connectivity. “There are a lot of great solutions for that first mile, last mile,” says Danville Mayor Newell Arnerich, noting the safety and convenience of being picked up at home and dropped off at a regional rail station.

Beyond these logistical debates, the commercialization of robotaxi services poses tough questions about accessibility. Advocates point out that they provide accessible transportation services to people who can’t drive. But as San José’s Madou notes, accessibility isn’t just about disability or age. “We’re trying to make sure emerging mobility technologies are within reach for people across the socioeconomic spectrum,” he says. “In San José, we have tools like our Equity Atlas to help companies identify communities that may have greater needs and fewer resources.”

In the three years since California green-lit commercial robotaxis, Waymo’s ridership has exploded, from 25,000 weekly trips in 2024 to almost 300,000 as of March 2026, according to the California Public Utilities Commission. (Waymo’s data suggests that number could be higher.) Waymo’s collision data shows a sharp drop in accidents in its four main markets. According to Sperling, Waymo now accounts for about a tenth of San Francisco’s ride-hailing market. Los Angeles County is at about 5% and increasing.

Even though scholars are grappling with technical safety riddles, few dispute the plunging collision stats and anecdotal evidence of robotaxis calming street traffic. A Waymo study found an over 90% reduction in both reported-injury collisions and airbag deployments compared to human-driven vehicles. “The safety promises for automation are very real, and they’re very real from a Vision Zero standpoint,” says William Riggs, a professor at the University of San Francisco School of Management and director of the Autonomous Vehicles and the City Initiative.

Some industry representatives acknowledge stubborn safety concerns, such as solo passengers who may find themselves in dangerous situations or unable to exit a vehicle due to malfunctioning equipment.

“A lot of people feel like AVs are safer because there is no driver in there, but you still have a human passenger,” Danielle Lam, Uber’s head of California policy, said at a recent University of San Francisco forum on robotaxis. “A couple of community groups in San Francisco asked me what happens if a woman is fleeing a violent situation and her abuser comes out and stops in front of the vehicle? Of course, the vehicle is going to stop because that’s what it is designed to do. This is something Uber spends a lot of time thinking about.”

The sharp ridership growth suggests a softening of public attitudes in some places, especially in areas with high concentrations of tech workers. However, overall public attitudes are still a mixed bag. Sperling cites a recent survey showing that 20% of respondents said they’d never ride in a robotaxi, and another 20% doubted they would. But as people become acquainted with the vehicles, their comfort with them increases. “The reality is that people are also just nervous and skeptical of new technologies.”

Danville Mayor Arnerich agrees, as he and other local officials discovered with an autonomous micro-bus trial. “We’ve done three test programs in Contra Costa County, and the group that embraced autonomous vehicles more than anybody was people over 65.”

“City infrastructure is not prepared.”

Complicating the picture is the degree to which these autonomous vehicles are actually autonomous. Waymo fleets are monitored remotely by human operators who are supposed to respond to passenger requests or collisions. But operator call centers are often located overseas, and a single operator may be responsible for large numbers of vehicles. In 2024, California legislators stepped up their efforts to guarantee that such communication channels work more effectively with emergency responders, with new regulations due this summer.  

Some city transportation officials stress that the non-collision interactions between robotaxis and other road users will only become more significant as fleets grow. “As these vehicles are coming in, we’re trusting the DMV and the federal DOT and CPUC to certify the vehicle safety and the system safety,” says San José’s Madou. “But at the city level, what we care about is that the bicyclists feel more comfortable driving on the roadway, that pedestrians know how to interact with the systems, and that vehicles pose less risk to everyone else than they do today.”

Another topic of mounting concern for city transportation services officials: curb spaces for robotaxis waiting to collect passengers. Bern Grush, a Toronto transportation entrepreneur, founded a platform to help cities manage such spaces. He notes that cities have long designated zones on public streets for specific users — taxi stands, bike lanes, and high-occupancy vehicle lanes — and robotaxis should be no exception. “That is a completely missing interface,” he says. “City infrastructure is not prepared.”

Bayen, at UC Berkeley, points out that cities own the rights-of-way, the curb, and traffic lights — all funded by local taxpayers. “It is legitimate that they should retain some control over how these things are deployed,” he says, drawing an analogy to how cities regulate vehicle noise. “If some of the robotaxi services start to impact the quality of life or the service level that people are expecting from their cities, I think it’s legitimate that they would regulate, like any other new technology deployed.” 

In Australia, a recent study concluded that robotaxi companies and public agencies should work together to roll out the technology. Similarly, Henriette Cornet, an adjunct professor at the University of San Francisco, points out that the testing and deployment of autonomous vehicles in Singapore involved collaboration between the industry and state authorities. In the U.S., she observes, the deployment of such technologies is mostly market-driven. The risk, Cornet adds, is that “there is no guarantee that the public good at the end will be satisfied.”

With accelerating growth in robotaxi fleets and repeat customers, questions that were once theoretical are gaining greater salience — chief among them their impact on transit. Riggs and others see robotaxis as an important solution to the last-mile problem. “For me, as a researcher, it’s an opportunity for reinvention, or re-imagination, of what is the future of public transportation, given that public transportation hasn’t had a reinvention for 100 years,” he says.

State Senator Scott Wiener agrees. “There are people who are of the view that public transportation is obsolete, that AVs are going to eclipse it,” he said at a University of San Francisco robotaxi symposium this spring. “That is false. AVs can complement public transportation, particularly around last-mile or in areas that are not as well served with transit. But they do not replace it.”

As Daniel Sperling adds, “if it’s really a traditional robotaxi, as we’re seeing with Waymo, there are legitimate concerns about increased traffic. But if you have the pooled services, you have the potential to provide huge benefits — much lower costs and much enhanced mobility for young, old, physically and economically disadvantaged, and other car-less travelers.”

Danville Mayor Arnerich, for his part, hopes that the technology, especially if it is eventually deployed in the form of autonomous micro-buses that can be hailed from any location, will help local and regional transit agencies struggling to serve declining passenger bases. “They are the most important partners,” he says, “and they have the opportunity to do it.”