Tesla Presents Misleading FSD Safety Data to EU Regulators

Per an exclusive Reuters investigation published June 15, 2026, Tesla submitted self-published safety statistics about its Full Self-Driving (FSD) system to regulators in the Netherlands and Sweden as part of efforts to secure wider European approval. Reuters reports the submitted materials claimed FSD-equipped cars travel more than seven times farther between crashes and can be up to ten times safer than the average human driver. A slide deck referenced in correspondence also included a claim that FSD could have prevented 32,000 deaths and 1.9 million injuries, according to Reuters.

Per Reuters, independent traffic-safety researchers flagged multiple methodological problems in the numbers Tesla supplied. Reuters documents and reporting identify at least two recurring issues: Tesla compared crash rates that counted only incidents severe enough to trigger airbag deployment in FSD-equipped vehicles against broader U.S. crash statistics that include many less-severe incidents, and Tesla benchmarked its relatively new fleet against the average U.S. vehicle fleet, which is materially older and less likely to include modern passive-safety features. Those comparisons, researchers told Reuters, materially inflate the appearance of an FSD-specific safety advantage.

Industry-pattern observations: Comparing non-equivalent severity buckets and mismatched fleet cohorts is a common pitfall when organisations present operational-safety claims. For practitioners, such errors reduce reproducibility and make independent validation difficult because they conflate vehicle hardware age, safety feature set, and driving-assistance usage into a single aggregate metric. Robust safety benchmarking for advanced driver assistance systems typically requires matched-cohort analyses, common severity definitions, and transparent exposure metrics such as miles driven under comparable operating conditions.

Reuters reports the Dutch road authority RDW emphasized it "does not rely on marketing claims or external statistics" and performs its own tests, analyses and verifications. Reuters also reports Swedish authorities told the news agency they "look beyond headline figures" and would not base decisions on aggregated claims alone. Reuters further reports RDW began a formal approval process after Tesla approached the agency in late 2024 and, following more than a year of testing and discussion, RDW approved FSD for use in the Netherlands in April 2026 and is seeking EU-wide approval on Tesla's behalf.

Editorial analysis: The episode highlights a broader regulatory friction point between self-reported operational metrics and independent safety assessment. Observers and regulators increasingly demand transparent, auditable evidence for safety claims in automated-driving technologies. For the AI and data-science community, this story underscores the stakes of data provenance, pre-specification of evaluation metrics, and the need for public, reproducible analyses when safety is asserted at scale.

• •Whether RDW or other EU bodies publish technical evaluation results that reproduce or rebut the submitted numbers.
• •Any regulator requests for additional, raw exposure data or for matched-cohort analyses.
• •Public responses, clarifications, or additional documentation from Tesla; Reuters reporting indicates regulators examined Tesla-provided materials but also conducted independent testing.

Per Reuters and corroborating coverage in outlets including Automotive World and Electrek, the specific numbers and methodology Tesla presented to some European regulators have been questioned by independent researchers and described in reporting as misleading. RDW and Swedish authorities told Reuters they apply independent verification rather than relying on headline claims.