Waymo's Robotaxi Safety Claims Put to the Test in Atlanta

Let’s face it – a car that doesn’t have a driver behind the wheel can feel a little spooky. You might wonder whether the robot will swerve, stall, or just ignore a frantic hand‑wave. That’s the vibe many Atlantans have had since Waymo’s self‑driving taxis rolled onto 17th Street last year.

Waymo, however, is pretty confident that its fleet is actually safer than the average human behind the wheel. The company points to a handful of peer‑reviewed studies – ten, to be exact – and a fresh crunch of crash numbers that look, at first glance, pretty impressive.

According to senior researcher Kristofer Kusano, Waymo’s vehicles suffered 95 % fewer “serious injury‑or‑worse” crashes than a typical driver covering the same distance. He defines a serious‑injury crash as one that lands a person in the hospital. The comparison, he says, is made over identical mileage in cities where Waymo has logged enough miles to be statistically meaningful.

From September 2020 (the date of Waymo’s first recorded crash) through June 2026, the sensor‑packed cars racked up more than 271 million miles. That’s a lot of data – every sensor ping, every sudden brake, every tiny fender‑bender gets logged. In contrast, human‑driver crash reports are spotty at best, especially for minor bumps that don’t make the police blotter.

To level the playing field, Waymo’s team filtered the data by vehicle type (mostly Jaguar sedans), road type (primarily surface streets) and geography (the handful of cities where they operate: San Francisco, Los Angeles, Phoenix, Austin and, yes, Atlanta). After that cleanup, the math showed not just fewer serious injuries, but also an 82 % drop in airbag‑deploying crashes.

The numbers get even more striking when you look at vulnerable road users. Waymo taxis were involved in 93 % fewer pedestrian injury crashes, 86 % fewer cyclist incidents and 82 % fewer motorcyclist injuries compared with human drivers. In Atlanta alone, where the fleet has logged about 8.6 million rider‑only miles, the injury‑crash rate was a staggering 88 % lower than the human‑driver baseline.

Waymo isn’t the only one bragging. The Insurance Institute for Highway Safety (IIHS) released a July 2026 study that echoed many of the same findings. According to that report, Waymo’s police‑reportable crash rate was 68 % lower than that of human drivers, rear‑ends were 91 % fewer, and even being rear‑ended happened 40 % less often for the robotaxis.

Still, the technology isn’t flawless. Some Atlanteans have reported odd moments – a Waymo cruising past a stopped school bus, another navigating around an active crime scene, or a vehicle getting stranded when it rolled into flood‑water. Those incidents remind us that a car that can’t shout “Hey! Hey! Heeeeeey!!!” can also be hard to communicate with in a pinch.

Waymo acknowledges the gaps and says it’s constantly tweaking its algorithms, learning from every mile. The company cites the World Health Organization’s “Decade of Action for Road Safety,” which aims to halve global crash injuries by 2030, as a big part of its motivation.

Bottom line? The data we have so far paints a picture of robotaxis that, at least on paper, keep people safer than the average human driver. Whether that translates into everyday peace of mind for Atlanta commuters is a story that’s still being written.