Leaked internal data and a landmark $243 million verdict reveal Tesla hid thousands of fatal autonomous driving incidents while continuing unregulated testing.

The RTS investigation draws on a massive Tesla data leak documenting over 2,400 sudden-acceleration complaints and more than 1,000 accidents, many marked "unresolved" years after Tesla knew of systemic AI "hallucinations" (rts.ch, 2026). This aligns with the NHTSA's January 2023 recall of 2 million vehicles over Autopilot defects and NTSB reports on 20+ fatal crashes from 2016-2022; original coverage underplayed the deliberate suppression pattern and its direct tie to sustained beta testing on public roads without mandatory incident reporting (nhtsa.gov; ntsb.gov).

What RTS missed was the parallel to documented AI failures in other domains: recovered "black box" data in the Naibel Benavides case proved the system detected pedestrians yet issued only a last-second alert, echoing hallucination errors catalogued in a 2024 MIT Technology Review analysis of autonomous systems and a 2025 AI Now Institute report on corporate accountability gaps that allow firms to treat fatalities as proprietary data (technologyreview.com; ainowinstitute.org).

The upheld February federal verdict exposes persistent regulatory shortfalls at NHTSA and DOT, which have relied on voluntary disclosures rather than audited, real-time reporting; synthesizing these sources shows Tesla's approach fits a broader pattern across AI developers where transparency deficits delay meaningful oversight and endanger non-consenting road users.