⚠️This fact has been debunked
This fact is completely outdated and inaccurate. By May 2015, Google's self-driving car project had already reported 14+ accidents (11 rear-end collisions), not just two. The program began in 2009, and by 2025, Waymo (Google's autonomous vehicle subsidiary) has been involved in 696 reported accidents from 2021-2024 alone. While it's true that most accidents were caused by human drivers (not the autonomous system), the 'only two accidents' claim is factually false.
The Google driverless car has only encountered two accidents. The first it was rear ended at a stop sign, and the second was when a human was behind the wheel.
The Myth of Google's Near-Perfect Self-Driving Record
You've probably heard the claim: Google's self-driving car only had two accidents—one rear-ending at a stop sign, and another when a human took the wheel. It sounds impressive, almost too good to be true. That's because it is.
By May 2015 alone, Google had already reported 14 accidents, with 11 of those being rear-end collisions. Fast-forward to 2025, and Waymo (Google's autonomous vehicle subsidiary) has logged 696 reported accidents between 2021 and 2024. So where did this "only two accidents" myth come from?
The Real Safety Story
Here's what's actually remarkable: in the overwhelming majority of these accidents, the self-driving car wasn't at fault. Google's data from 2015 showed that "not once has the self-driving car been the cause of the collision." The cars were rear-ended at traffic lights, hit by drivers rolling through stop signs, and sideswiped by inattentive humans.
One 2015 incident perfectly captures this pattern. A Google Lexus, traveling in autonomous mode, was rear-ended at low speed while obeying a red stop sign. The human driver behind apparently got confused by a green light for a different lane and plowed into the back of the robot car.
How Safe Are They Really?
When you dig into Waymo's modern safety data, the numbers are genuinely impressive—just not in the way the myth suggests. Swiss Re's analysis found that Waymo vehicles have:
- 92% fewer bodily injury claims than human drivers
- 88% fewer property damage claims over 25 million miles
- 79% reduction in airbag-triggering crashes
- 92% fewer crashes injuring pedestrians
Waymo estimates that typical human drivers would have caused 159 airbag-deploying crashes over the same 96 million miles where Waymo's cars had only 34 such incidents.
Why the Confusion?
The "two accidents" claim likely stems from early, limited reporting when the technology was still in its infancy. It may have referred to a specific vehicle or time period, but it was never representative of the entire program. The self-driving project began in 2009, and by the time this fact started circulating, dozens of incidents had already occurred.
What's fascinating is that even when Waymo vehicles crash, fault analysis consistently points to human error. In one six-month period, Waymo vehicles were involved in 45 crashes causing injury or triggering airbags. Of these, 24 occurred when the Waymo wasn't even moving, and 7 more were straightforward rear-endings.
So while the "only two accidents" claim is demonstrably false, the underlying premise—that autonomous vehicles can be safer than human drivers—appears to be supported by data. The real story isn't about a mythical perfect record. It's about a technology that's already outperforming humans, one rear-ending at a time.