Waymo's corrected NYC safety data: 62% fewer injury crashes and 87% fewer serious-injury/fatal crashes vs. for-hire vehicles
Hunter College retracted flawed July 2026 Schwartz report; corrected analysis shows Waymo 62% fewer injury crashes, 87% fewer serious/fatal vs NYC for-hire vehicles — evidence via retraction, not regulation
Headline Finding
A corrected Hunter College analysis — published after Waymo identified significant methodological flaws in the original — found that Waymo vehicles recorded 62% fewer crashes resulting in injury and 87% fewer crashes resulting in serious injury or death compared to New York City for-hire vehicles. The figures represent the most concrete safety benchmark yet introduced into the NYC AV policy debate, and their emergence through a retraction-and-correction cycle rather than through a stable regulatory evidence process underscores the governance clock’s lag behind the deployment clock.
Detail
The original Hunter College report, released in July 2026 under former NYC traffic commissioner Sam Schwartz, claimed Waymo vehicles were less safe than New York City for-hire vehicles. After Waymo identified significant flaws in the analysis, the authors retracted the study. A revised version, released a few weeks after initial publication, found Waymo taxis had “62% fewer crashes resulting in an injury and 87% fewer crashes resulting in a serious injury or death when compared to New York City for-hire vehicles.” Schwartz himself, in a separate interview, stated his personal belief that AVs “will prove to be safer than the conventional driver of general traffic” and that “professional drivers do have a lower crash rate per mile” — acknowledging that AVs are likely to surpass even that benchmark “at some point.” The corrected figures now constitute the operative safety evidence in the NYC robotaxi debate, having entered the public record through a contested, non-regulatory process.
Governance Implication
- The technology/deployment clock: the corrected data — 62% fewer injury crashes, 87% fewer serious-injury or fatal crashes versus NYC for-hire vehicles — represents a material safety-frontier advance that, if sustained at scale, would justify accelerated deployment consideration; the governance clock has not yet produced a regulatory framework in NYC capable of evaluating or acting on such data systematically.
- The legitimacy lever is doubly exposed: the data’s credibility is undermined by the retraction process, meaning even strongly favourable safety figures cannot anchor policy without an independent, institutionalised evidence-review mechanism — a gap cities must close before multi-operator AV markets open.
- For performance-based permitting, these figures set a potential benchmark: a city could require all AV operators to maintain injury-crash rates at or below the corrected Waymo threshold as a condition of continued operation, converting a contested research finding into a durable regulatory instrument.
- The episode signals that the pacing problem (chapter 1) is acute in NYC: the absence of a standing, credible safety-data infrastructure means the public debate is being shaped by retracted academic papers and science-fiction novels rather than by regulatory evidence — a condition that favours neither operators nor the public.