The claim
“Robotaxis are safer than human drivers.”
Waymo says that, over the same distance in the cities where it runs, its cars had far fewer crashes than an average human driver. Waymo Safety Impact page, data through June 2026
Peer-reviewed and insurer studies find Waymo crashes far less than human drivers where it runs; that does not cover every robotaxi company.
01The evidence
Independent
A study in the journal Traffic Injury Prevention, written by Waymo researchers, looked at 56.7 million driverless miles up to January 2025 and found Waymo had a lower crash rate than human benchmarks, at a level unlikely to be chance.
Kusano et al., Traffic Injury Prevention, 2025 (Waymo research page)
Independent
A study by insurer Swiss Re with Waymo, using insurance claims from 25.3 million driverless miles, found 88% fewer property damage claims and 92% fewer injury claims than human drivers.
Company or forecast
Waymo says that, over the same distance in its cities, its cars had 95% fewer crashes causing serious injury or worse than an average human driver.
Independent
Electrek's analysis of crash reports Tesla filed with US safety regulator NHTSA found Tesla's Austin robotaxis had 14 crashes in about 800,000 miles, roughly one every 57,000 miles, nearly all with a safety monitor on board.
Independent
Tesla's own figures say a typical human driver has a minor crash every 229,000 miles, so its robotaxis crashed more often by that yardstick.
Independent
NHTSA opened an investigation after Waymo cars passed stopped school buses; the Austin school district documented 19 such cases.
Independent
Waymo recalled about 3,800 robotaxis in the US after finding they could drive into flooded roads.
What we don't know. The main studies were written by Waymo or with Waymo, using its data. Few robotaxi companies besides Waymo publish crash rates that can be compared with human drivers.