Photo by Andri Aeschlimann on Unsplash
By Olivia “Biv” McNamara
As Waymo expands its driverless taxi service across the country, experts disagree over whether its lower crash rates prove that self-driving cars are safe enough for widespread use.
Waymo’s service has grown quickly. According to TechCrunch, the company went from providing 50,000 paid rides per week in May 2024 to 500,000 per week in March 2026.
As more driverless taxis take to the road, supporters point to lower crash rates as evidence that self-driving cars can be safer than vehicles driven by humans. Others say there are still situations on the road that the cars may not be prepared to handle.
Waymo says its cars are safer than human drivers. According to the company’s latest safety data, its vehicles had 94% fewer crashes resulting in serious or fatal injuries and 82% fewer crashes involving any injury than human drivers.
Independent research has also found that Waymo cars are involved in fewer crashes. A July 2026 study from the Insurance Institute for Highway Safety found that Waymo vehicles had 68% fewer crashes per mile than human drivers.
Steven Shladover, a research engineer at the University of California, Berkeley, who studies automated transportation, said, “Waymo in particular has done a more thorough job in assuring the safety of their system than any of the other companies.”
He also said, “I think that they have reached the point of showing that they are safe enough for use in the places where they already have a lot of experience.” Shladover said those places include surface streets in San Francisco, Phoenix, and Los Angeles.
Amber Rollins-Reis, executive director of Kids and Car Safety, a national nonprofit dedicated to vehicle safety, said that the technology has great potential to prevent crashes by eliminating human error as a factor.
However, Rollins-Reis also said, “Widespread deployment should be guided by rigorous, transparent, independent safety data and strong federal oversight, with safety proven before — not after — these vehicles are deployed at scale.”
Philip Koopman, a professor emeritus at Carnegie Mellon University who studies autonomous vehicle safety, said lower crash rates do not tell the whole story. In his essay, “How Safe Is Really ‘Safe Enough’ for Autonomous Vehicles?,” Koopman examines whether current crash data is enough to determine their safety.
“There is more to safety than serious crashes,” said Koopman. “Waymo simply does not have enough data to know how fatal crashes will turn out. Their claim of already saving lives is hope and hype.”
Koopman also pointed to problems involving stopped school buses, floodwaters, construction zones, wrong-way driving, and interpreting traffic direction gestures. He said these problems need to be addressed as Waymo expands.
Peter Kurdock, general counsel for Advocates for Highway & Auto Safety, said that crash statistics alone are not enough to evaluate the technology, especially as driverless cars encounter more situations on public roads.
“To properly assess the safety of these operations, additional performance data beyond crashes is needed, considering recent events such as Waymo vehicles traveling on light rail tracks and passing school buses,” Kurdock said.