What will the autonomous vehicle startup Waymo find once it starts mapping the streets of Boston, as it plans to do this month? Regulatory hurdles, rotaries and roads riddled with potholes, for a start.
I’ve been talking with people who have experience testing autonomous vehicles (AVs) in Boston. Two companies with MIT roots have used the Seaport neighborhood for testing in the past few years: Optimus Ride and Motional. Neither ultimately launched a taxi or ride-sharing service here, and all of the testing was done with a human “safety driver” in the vehicles to take over in situations where the technology ran into problems.
Here’s what the veterans of Optimus Ride and Motional regard as the most significant challenges Waymo will face in trying to offer a truly driverless ride-share service in Boston, and the odds that it will be launched by next year. Waymo is a subsidiary of Google’s parent company, Alphabet, and its development of self-driving vehicles traces back to 2009.
LISTEN TO ‘KIRSNER ON INNOVATION’ ON WBUR HERE
1. Boston winters
What do the cities where Waymo currently operates have in common? Snow isn’t a regular feature of the weather forecasts in L.A., Austin, Phoenix or San Francisco. Waymo has reportedly tested its vehicles near Lake Tahoe, California; in Michigan; and upstate New York. The next two announced launch cities are Miami and Atlanta.
Snow can be a major challenge for autonomous vehicles trying to “see” the world with sensors, just as it is for human eyes. Lane markings get covered, and parked cars covered with snow can look like part of the background.
Karl Iagnemma, the former CEO of Motional, explains: “Heavy snow … can cause the sensors to hallucinate and perceive obstacles that actually aren’t there.” That company eventually moved its testing operations to Las Vegas, which has “almost perfect weather year-round.” (As part of company cutbacks last year, a taxi service Motional had operated there, with human safety drivers in the front seat, was shut down in 2024. The carmaker Hyundai owns a majority stake in Motional, and the company is headquartered in Boston.)
“There’s almost no significant amount of autonomous vehicle testing in cold weather,” Ryan Chin, former CEO of Optimus Ride, said. “Most of the pilots have been in Arizona, Nevada and California, where the weather is great.”
One possibility Chin and others mentioned is that Waymo cars could simply stop operating in certain weather conditions.
2. The most unpredictable vehicle on city streets is …
AVs have gotten pretty good at predicting what another car or a pedestrian is going to do, Chin said. But cyclists? They’re tough to track and predict, he said, because they travel relatively fast, and they don’t always follow the rules of the road. Scooter riders and delivery drivers on gasoline-powered scooters are in that same category.
“In general, those guys don’t follow the rules the most,” he said, zipping down one-way streets or moving from sidewalk to the road. AVs need to avoid hitting them.
3. Paved cow paths
Waymo cars may be able to summit San Francisco‘s hills and endure Phoenix’s blast furnace temperatures, but Boston’s ancient street layout? It may be quaint for tourists, but it’s an Olympic-level challenge for AVs.
Jenny Berlin, an Optimus Ride co-founder, noted that the core of Boston “was designed for horses and carriages,” making it “an older city than the other cities that Waymo‘s been in.”
Chin piled on: “We have five-way intersections, six-way intersections. A lot of the cow paths from the British [colonial era] led to the street grid we have. It’s very, very complicated here.”
Not to mention rotaries.
Those challenges could mean that Waymo vehicles try to route around certain areas of the city, Chin said.
4. AVs can’t be aggressive … but Boston drivers can
Sending rule-following AVs into the wild world of Boston driving may be a bit like tossing a Mozart fan into a mosh pit.
“For a while, I think Massachusetts was rated as the state with the worst road rage,” Berlin said, adding, “I don’t know, given Massachusetts drivers’ history, if they’re going to be as patient” with AVs that stop abruptly, hesitate or need remote operator intervention.
And Basak, formerly the chief product and design officer at Motional, reminded us that when AVs follow the rules — even in basic instances, like not pulling into a lane of oncoming traffic to move past a stopped delivery vehicle — it can frustrate everyone else.
“They’re not going to want to break the law, because nobody wants to give more attention to an AV than there already is,” she said. “So they’ll do everything properly.”
Boston entrepreneur Gary Ambrosino used Waymo‘s ride-share service in the Phoenix area recently, but “the streets are all straight and the drivers are not crazy,” he said. “I am looking forward to seeing Waymo in Boston … driving around the financial district and on Storrow Drive. If the rider survival rate is higher than driving yourself in Boston, then I’m convinced.”
5. Regulatory hurdles
While the city and state began mapping out the rules for testing AVs in Boston as early as 2016, no company has yet jumped through the hoops to get regulations changed to allow fully autonomous driving, without a human safety driver.
“At [the time we were testing], there was no provision for removing the safety driver,” Chin said. He and others say it’s not clear how long that will take Waymo.
Boston City Hall officials didn’t respond to numerous requests for comment.
6. An injury or fatality could freeze the roll-out
Waymo‘s website boasts that its technology is already “the world’s most experienced driver” after traveling more than 50 million miles, and early peer-reviewed research has already begun to show a reduction in crashes and injuries, relative to human drivers.
But even if an AV is statistically safer than a human driver, anytime it makes a mistake and injures a person is likely to be met with outrage, and potentially calls to shut the services down. Chin put it bluntly: “There is this general fear of the unknown,” and even if AVs are statistically safer than human drivers, people will place “a higher level of expectation on this genius machine,” expecting complete perfection.
“The scientific argument should be … that AVs should be tested to the point where they are statistically better than humans,” Chin said. “Otherwise, it’s not worth the effort to do this, because we can continue to kill people the way we normally do, which is, to drive.”
In Massachusetts, according to Department of Transportation data, there were 2,681 accidents that caused a fatality or serious injury in 2024. In San Francisco, an early test city for both Waymo and another AV operator, Cruise, a small handful of accidents have involved cyclists or a pedestrian in the roadway who had already been hit by a human driver.
In 2018, an Uber test vehicle with a safety driver at the wheel struck and killed a pedestrian in Tempe, Arizona; that individual was pushing a bicycle across the road, outside of a marked crosswalk, and the safety driver was found to be watching television on her phone.
7. Just a data-gathering detour?
Waymo‘s arrival in Boston has generated buzz, but Chin was skeptical of over-interpreting the company’s mapping exercise: “This is not necessarily … a commitment to launch,” he said. “It certainly opens the door … but you can’t look at mapping as a commitment.”
Jenny Berlin echoed that sentiment, suggesting that this could simply be early reconnaissance.
“They have to understand the different environments they’re going to go to,” she said. The company likely has a checklist that includes things like the cost of facilities, how much it will cost to hire employees, how much time it will take to work through the regulatory issues and whether there are enough “early adopters” willing to hop into an AV here. “You need to look at the whole ecosystem, in terms of meeting [Waymo‘s] requirements” for a launch, she said.
Waymo spokesperson Sandy Karp said via email that the company doesn’t have any announced plans to launch service for riders in Boston.
“We continue to focus on safely and responsibly advancing our autonomous technology through short visits like this in various cities around the U.S.,” Karp wrote.
I asked each of the AV veterans if they thought Waymo would be offering rides to paying customers one year from now.
Chin said he thought the chances were 50-50, “and the caveat is that it’s probably going to be certain time frames, and only in certain sections of the city,” he added.
Both Berlin and Iagnemma were slightly more enthusiastic than that: “I’d say that it’s more likely than not that they are operating in Boston one year from now,” Iagnemma said.
Ozer was the most bullish of the group, putting the odds at 80-20, “They’ll make it happen,” she said.
