A Reuters investigation published May 28, 2026, alleges Tesla uses a flawed methodology for its Full Self-Driving safety statistics. The report cites interviews with nine former Tesla data labelers and one former self-driving engineer. These sources claim that Tesla’s own AI trainers do not trust the technology and that it struggles with basic maneuvers.
The investigation highlights that Tesla inflates safety claims by comparing its airbag-deployment crashes to federal data that includes all tow-away accidents. This specific discrepancy reportedly increases the claimed safety level by a significant factor. These findings cast doubt on the central promises driving Tesla’s high market valuation and future robotaxi ambitions.