Timothy B. Lee has been writing full-time about AI on his Substack, Understanding AI, since 2023. In his latest post, he analyzed Waymo's recent crash data and determined that Waymo robotaxis drive much better than human drivers. This may seem surprising if your perception of autonomous vehicles is based upon the many, many "full-self-driving" Tesla crashes.
Lee analyzed crash data from July 2024 and February 2025 and found 38 crashes with an (alleged) injury or airbag deployment. Only one of those was unquestionably the fault of the Waymo vehicle, and it was a unique situation, involving a plastic crate and a scooter. He could not determine the party at fault with certainty in three of the accidents. Even if you give the human drivers the benefit of the doubt, Waymo would only be at fault for at most four of 38 crashes.
When Lee looked at the per-mile crash rate, the numbers were even more dramatic. Through December 2024, Waymo vehicles drove 44 million miles in Phoenix and San Francisco.
Using human crash data, Waymo estimated that human drivers on the same roads would get into 78 crashes serious enough to trigger an airbag. By comparison, Waymo's driverless vehicles only got into 13 airbag crashes. That represents an 83 percent reduction in airbag crashes relative to typical human drivers.
In a similar comparison of accidents with injuries, Waymo maintained an 81 percent reduction. An analysis of insurance data, where a third party determines fault, shows that Waymo's autonomous vehicles reduced bodily injury claims by over 90% compared to human drivers. None of this is to say there aren't issues with Waymo vehicles, but there is a clear safety advantage.
Previously:• Multiple workers required to observe single puddle in San Francisco
• Waymo sues accused robotaxi vandals
