Waymo is harnessing insurance data related to autonomous vehicles to strengthen its safety argument.

Waymo is leveraging insurance data to strengthen its safety argument for self-driving cars. The Alphabet-owned company asserts that its autonomous vehicles result in fewer personal injury and property damage claims compared to human drivers.

This initiative follows the collaboration announced last year between Waymo and the insurer Swiss Re and aims to demonstrate the lower accident frequency and property damage associated with driverless vehicles in comparison to their human-operated counterparts.

Waymo’s analysis arrives at a time when self-driving vehicles are facing heightened scrutiny, particularly after California regulators expanded the operational scope of robotaxi companies in San Francisco, despite concerns from city officials about traffic disruptions and delays to emergency responders caused by autonomous vehicles.

The study conducts a comparative analysis between Waymo’s liability claims data and those of human drivers, taking into account mileage and zip code factors. Waymo maintains that the data, based on Swiss Re’s dataset comprising over 600,000 claims and 125 billion miles, is highly robust and statistically significant. The company intends to submit and publish this research in a peer-reviewed journal.

According to the comparisons, Waymo’s autonomous vehicles outperform human-driven ones in terms of safety. In fully autonomous mode without a safety driver, Waymo’s vehicles reduced bodily injury claims by 100%, compared to Swiss Re’s human baseline of 1.11 claims per million miles. This assessment was based on over 3.8 million miles of autonomous driving by Waymo in California and Arizona.

Similar positive results were observed for Waymo’s vehicles operating in autonomous mode with a safety driver. These vehicles reduced injury claims by 95% over 35 million miles, recording 0.09 claims per million miles versus the human baseline of 1.09 claims. Furthermore, Waymo reported a 76% decline in property damage claims, decreasing from 3.26 to 0.78 claims per million miles when compared to human drivers.

Notably, Waymo asserts that this represents the first instance where liability claims data has been employed to directly compare the safety performance of autonomous and human drivers.

Mauricio Peña, Chief Safety Officer at Waymo, emphasized the role of the Waymo Driver in enhancing road safety and reducing injuries in cities where it operates. He underlined the significance of this study, providing compelling evidence of the Waymo Driver’s impact on injury reduction.

The data holds importance due to the ongoing debate surrounding the safety of autonomous vehicles in comparison to human drivers. Companies like Waymo argue that self-driving cars are essential to address the high incidence of traffic fatalities, approximately 40,000 annually in the US. They highlight that autonomous vehicles don’t experience intoxication, fatigue, or distraction, minimizing the human errors responsible for accidents and fatalities.

However, the uncertainty regarding the safety of driverless vehicles persists, primarily due to the limited number of autonomous vehicles on the road compared to human-driven vehicles, resulting in less data available for conclusive assessments. Human-driven vehicles typically travel close to 100 million miles between fatal crashes, necessitating the accumulation of hundreds of millions of miles from autonomous vehicles to enable more meaningful safety comparisons.

In the interim, Waymo has been actively working to bridge the knowledge gap regarding self-driving cars by submitting its own data for analysis. The company published two scientific papers last year, one analyzing response times in imminent crash scenarios and the other presenting a novel methodology for evaluating autonomous driving systems’ crash avoidance performance. Waymo has also conducted simulations of real-world fatal crashes in Arizona, revealing that replacing one of the vehicles involved with its autonomous counterparts could significantly reduce fatalities.