Tesla's FSD ‘32,000 lives saved’ claim challenged by experts
Tesla is using average American vehicle as comparison point, although it is an older model with fewer safety technologies available
Tesla has been presenting self-published safety statistics to European regulators that independent traffic researchers say amount to misleading marketing, and the figures were used to help secure approval for its Full Self-Driving (FSD) system in the Netherlands, with EU-wide approval now being pursued on that basis.
A Reuters investigation published last month identified several invalid comparisons underpinning Tesla's safety claims.
While the company has claimed that FSD-equipped cars are at least ten times safer compared to humans, the review revealed that Tesla gets such results by comparing its airbag deployment crashes per capita ratio to the US standard that consists of much less serious crashes.
Tesla is also using the average American vehicle as a comparison point, although it is an older model with fewer safety technologies available.
A PowerPoint slide presentation provided by Tesla policy manager Ivan Komusanac to Swedish officials in April included a claim stating FSD could have prevented 32,000 deaths and a further 1.9 million injuries.
According to Reuters, this statistic is based on the very unrealistic assumption that all vehicles used in America, including trucks and injury-prone motorcycles, will be replaced with Teslas with FSD and that each one is at least seven times more secure.
"Give the data to a university, have it independently verified by a qualified researcher, and then let's talk," said the spokesperson for the watchdog group European Transport Safety Council Dudley Curtis.
In late 2024, Tesla contacted the Dutch road regulation body Red Cell Distribution Width (RDW) and shared a link to its safety report that "leads to safer roads". In April, FSD got the go-ahead from RDW to operate in the Netherlands, with the RDW working to approve its operation in all of Europe on Tesla's behalf.
RDW did not depend on marketing claims or any external statistics but tested the system itself, yet it would not confirm whether or not they tested the statistics from the US.
Following the Dutch approval of FSD, Komusanac contacted Swedish officials and asked for similar approval, sharing the exact presentation with the disputed statistics.
-
AI's new metric: Intelligence per dollar, not just smarts
-
How did OpenAI miss its own agent going rogue?
-
What happens if the internet goes down for good?
-
Is sharing health data with ChatGPT worth the risk?
-
Why is ChatGPT down? OpenAI confirms widespread service disruption
-
How Samsung Electronics $200 billion Broadcom deal could reshape the AI chip market
-
How to preorder new Samsung Galaxy Z Fold 8, Z Flip 8, Z Fold 8 Ultra
-
X removes 42,000 AI chatbot accounts, wins praise from users