Fact Check: Waymo are safer then Humans

Are Waymo's Autonomous Vehicles Safer Than Human Drivers?

Introduction

The claim that "Waymo is safer than humans" refers to assertions made by Waymo and various studies suggesting that their autonomous vehicles (AVs) demonstrate significantly better safety records compared to human drivers. This claim has garnered attention as the technology behind self-driving cars continues to evolve and integrate into everyday transportation. However, the validity of this assertion requires careful examination of the available evidence, methodologies, and potential biases in the sources.

What We Know

1. Waymo's Safety Performance: Waymo has published studies indicating that their autonomous driving system (ADS) has a lower crash rate compared to human drivers. According to Waymo, their vehicles have achieved an 85% reduction in crash rates involving injuries, with a reported incidence of 0.41 crashes per million miles compared to 2.78 for human drivers 2.

2. Swiss Re Study: A study by Swiss Re, a global reinsurance company, claims that Waymo's AVs are significantly safer than human drivers. This study analyzed auto liability claims related to 25.3 million miles driven by Waymo and found an 88% reduction in property damage claims and a 92% reduction in injury claims when compared to human benchmarks 34.

3. Comparative Analysis: Additional reports indicate that Waymo's vehicles have logged over 50 million miles, with significantly fewer incidents compared to human drivers. For instance, Waymo's vehicles reportedly experienced only 13 airbag crashes, suggesting an 83% reduction in such incidents compared to typical human driving 9.

4. Broader Context: The comparison of AVs to human drivers is complex, as it involves various factors including driving conditions, types of roads, and the nature of the incidents. The studies often rely on specific methodologies that may not account for all variables affecting driving safety.

Analysis

Source Evaluation

• Waymo's Own Reports: The data from Waymo 12 is self-reported and may carry inherent bias, as the company has a vested interest in promoting its technology as safer. While the methodologies used in their studies are detailed, independent verification is crucial for establishing credibility.

• Swiss Re Study: The Swiss Re study 348 is more credible due to its independent nature. However, it is essential to consider that Swiss Re operates within the insurance industry, which may influence its perspective on AV safety due to potential impacts on liability claims and insurance models.

• Media Reports: Articles from sources like NBC Bay Area 6 and Ars Technica 9 provide summaries of these studies but may not delve deeply into the methodologies used or the potential limitations of the data. While they can offer valuable insights, they should be approached with caution, as media outlets may prioritize sensational headlines over nuanced analysis.

Methodological Concerns

• Data Collection: The methodologies employed in these studies often involve comparing specific metrics (e.g., crash rates per million miles). However, the context of these miles driven—such as urban vs. rural settings, traffic conditions, and the presence of pedestrians—can significantly affect safety outcomes.

• Benchmarking Against Human Drivers: The benchmarks used for human drivers are critical. If the human baseline is derived from a limited or biased dataset, it could skew the results in favor of AVs. The reliability of the human benchmark data from Swiss Re, which includes over 500,000 claims, appears robust but warrants further scrutiny regarding its representativeness of all driving conditions 4.

Conflicts of Interest

• Industry Influence: Waymo's position as a leading AV developer means that its findings may be viewed as promotional rather than purely scientific. The potential for conflicts of interest exists, particularly in how data is presented and interpreted.

• Insurance Perspectives: Swiss Re's involvement in the study could reflect an interest in promoting AVs as a means to reduce liability claims, which may influence the framing of their findings.

What Additional Information Would Be Helpful?

• Independent Verification: More independent studies that replicate Waymo's findings using different methodologies and datasets would enhance the credibility of the claim.

• Longitudinal Studies: Long-term data on the performance of AVs compared to human drivers across various conditions would provide a more comprehensive understanding of safety implications.

• Detailed Methodologies: Access to the full methodologies used in the studies, including how human benchmarks are established and the specific conditions under which data was collected, would allow for a more thorough evaluation of the claims.

Conclusion

Verdict: Partially True

The claim that Waymo's autonomous vehicles are safer than human drivers is partially true based on the evidence presented. Waymo's own reports and the Swiss Re study indicate significant reductions in crash rates and claims when comparing their AVs to human drivers. However, the self-reported nature of Waymo's data raises concerns about potential bias, and the methodologies used in these studies may not fully account for all relevant variables affecting driving safety.

The complexity of comparing AVs to human drivers, alongside potential conflicts of interest from both Waymo and the insurance industry, adds further nuance to the claim. While the evidence suggests that Waymo's AVs may be safer in certain contexts, the limitations in data and methodology necessitate caution in drawing definitive conclusions.

Readers are encouraged to critically evaluate the information presented and consider the broader context when assessing the safety of autonomous vehicles compared to human drivers.

Sources

1. Comparative safety performance of autonomous and human drivers - Waymo. Link
2. Waymo significantly outperforms comparable human benchmarks over 7 million - Waymo Blog. Link
3. Waymo's AVs Safer Than Human Drivers, Swiss Re Study Finds. Link
4. New Swiss Re study: Waymo is safer than even the most advanced human - Waymo Blog. Link
5. Waymo robotaxis are safer than human drivers | GrowSF.org. Link
6. Waymo's robotaxis surpass 25 million miles, but are they safer than - NBC Bay Area. Link
7. Waymo Reduces Crash Rates Compared to Human - AV Industry. Link
8. How Waymo's AI-Driven Vehicles are Making Roads Safer. Link
9. After 50 million miles, Waymos crash a lot less than human - Ars Technica. Link
10. Comparison of Waymo Rider-Only crash rates by crash type to human - Taylor & Francis Online. Link