Reuters: Tesla's FSD safety claims are misleading marketing, not safety research (electrek.co)

A recent Reuters investigation has raised serious questions about Tesla's "Full Self-Driving" (FSD) safety claims, revealing that the company's statistics are founded on flawed methodologies. Interviews with former data labelers and experts indicate that Tesla's assertion of being "up to 10 times safer than human drivers" is significantly exaggerated—by as much as a factor of three. The investigation highlighted a critical comparison error where Tesla compared airbag-deployment crashes in its vehicles to broader federal data that included non-airbag incidents, resulting in misleading claims about the technology's safety. Such discrepancies undermine the credibility of Tesla's safety assertions, especially since leading traffic-safety researchers have labeled these statistics as more marketing than genuine safety analysis. Moreover, the report unveiled deep mistrust among Tesla employees towards the FSD system itself, with many data labelers expressing reluctance to use it for their own safety. Observations of FSD's failures in basic driving tasks, such as navigating around emergency vehicles and properly reacting to pedestrians, paint a troubling picture of the system’s current capabilities. Additionally, Tesla's practices of mapping robotaxi zones ahead of public launches, contrary to Elon Musk's claims of avoiding extensive local mapping, further illustrate the challenges facing the company in achieving true autonomous driving. As Tesla faces mounting regulatory scrutiny, these revelations stress the pressing need for transparency and reliable safety data in the progression of autonomous vehicle technology.