A Note on Tesla's revised safety report crash rates
DOI: 10.55329/llfl7748
archive: archived pipeline: cataloged verified
Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)
Summary
This short communication examines discrepancies in Tesla’s quarterly safety reports, which serve as the primary national source for crash rates involving Level 2 advanced driver assistance systems (ADAS). The author investigates changes Tesla made in January 2023, when the company revised historical data from June 2018 to December 2023. Prior to March 2021, Tesla categorized crashes into three groups: with Autopilot engaged, without Autopilot but with active safety features, and without either. In 2023, Tesla consolidated these into two categories: “using Autopilot technology” and “not using Autopilot technology,” while also adjusting for double-counting and excluding crashes that did not involve airbag or active restraint deployment. The analysis compares pre-revision and post-revision crash rates. For Autopilot crashes, the average change was a modest 3.6% reduction. However, the non-Autopilot category revealed significant anomalies. The author notes that the revised “not using Autopilot” crash rates closely matched the pre-revision rates for vehicles with *no* active safety features, rather than those *with* active safety features. This finding is counterintuitive because external research indicates that 93% of drivers keep active safety features, such as automatic emergency braking, enabled by default. Furthermore, Tesla’s own manuals state these features activate automatically at every ignition cycle and operate across nearly all driving speeds. To explain this discrepancy, the author evaluates three potential causes. First, a “universal quantification” approach—where a crash is classified as having active safety features only if *all* features are simultaneously active—is deemed impossible because different features have mutually exclusive operational speed ranges. Second, alternative classification schemes, such as excluding specific features or requiring multiple simultaneous features, are considered plausible but unverified. Third, the author analyzes the statistical weights of the combined averages. Calculations suggest that for the revised rates to hold, vehicles would need to have traveled approximately 6.3 times further without active safety features than with them in certain quarters. This contradicts the known prevalence of enabled safety features and statistical expectations regarding variance in large sample sizes. The study concludes that the most efficient explanation for the observed data is that Tesla undercounted or excluded a portion of crashes involving active safety features during the revision. The author argues that these discrepancies highlight a lack of transparency in Tesla’s reporting methodologies. Given the scarcity of independent ADS crash data, the paper emphasizes the need for clearer definitions and greater data transparency from manufacturers to allow researchers and regulators to accurately assess the safety of automated driving systems.
Provenance
The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| discover | success | DOAJ | — | — | 1 | 2026-06-17 |
| archive | success | unpaywall | — | — | 1 | 2026-06-25 |
| extract | success | cached | — | — | 2 | 2026-06-25 |
| clean | success | clean | — | — | 1 | 2026-06-18 |
| chunk | success | chunk | — | — | 1 | 2026-06-18 |
| embed | success | embed | Qwen/Qwen3-Embedding-8B | — | 1 | 2026-06-18 |
| promote | success | — | — | — | 1 | 2026-06-17 |
| summarize | success | llm | qwen3.6-27b-prismaquant | summ-v5 | 1 | 2026-06-25 |
| tag | success | vector_similarity | — | — | 6 | 2026-06-18 |
| verify | success | — | — | — | 1 | 2026-06-26 |
Summary generated by qwen3.6-27b-prismaquant on 2026-06-25; verification: verified.
Topics
Ranked by relevance to this paper. Hover a topic for its definition.
Information type
What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).
- Empirical Findings: crash risk outcomes