Tesla's relentless push toward autonomous driving has always balanced audacious marketing with hard data. In July 2025, the company touted Autopilot as "already 10x safer than the average US driver" in a widely shared X post, backed by fleet statistics showing one accident per 4.1 million miles driven with the system engaged, which far outpaced the US average of 0.38 million miles.
A Tesla using Autopilot technology today is already 10x safer than the average US driver pic.twitter.com/uhEpDkVuNE
— Tesla (@Tesla) July 24, 2025
Yet, recently, Tesla has quietly updated the owner’s manuals with software update 2025.38, removing “safer” from the longstanding description of Autopilot and FSD features. Autopilot’s description used to be:
“Autopilot refers to a suite of advanced driver assistance features that are intended to make driving safer and less stressful.”
However, the revised text now reads:
“Autopilot refers to a suite of advanced driver assistance features that are intended to make driving more convenient and less stressful.”
Gone is the word “safer,” and it’s been replaced with “more convenient.” That’s a huge change for a company that advertises and believes that Autopilot is 10x safer than a human.
Why the Change?
Historically, Tesla has leaned into safety as a core selling point for Autopilot and FSD. While Tesla isn’t wrong to use the term "convenient," which is a major selling point of Autopilot, removing the "safer" keyword seems to contradict what the company states and advertises outside the owner’s manual.
So why now? Tesla’s latest Vehicle Safety Report continues to champion FSD and Autopilot’s progress with safety. As of Q2 2025, vehicles using Autopilot logged accidents at a rate 10 times lower than the national baseline, based on over 9 billion global miles driven. This data, drawn conservatively from Tesla’s airbag deployment events or other critical incidents (excluding minor incidents), paints Autopilot as a statistical win, especially on highways, where it is most used.
Regulatory Heat
This change comes as government oversight into Tesla’s systems continues. The National Highway Traffic Administration (NHTSA) has probed Tesla’s driver assistance systems since 2021, but the investigations have been escalating. Just last month, the most recent investigation targeted 2.9 million vehicles over FSD or Autopilot-related incidents, including 14 crashes and 23 injuries, as well as minor traffic violations such as running red lights.
Was the manual change a legal requirement? Not explicitly. NHTSA hasn't mandated any wording tweaks as a result of its investigations to date. But it echoes calls from senators and Consumer Reports for the FTC to scrutinize "deceptive" claims, dating back to 2021, when Autopilot and FSD were accused of blurring the line between assistance and autonomy.
Critics Argument
Critics and regulators argue that Tesla’s fleet data, while extremely impressive, may undercount risks in urban scenarios, where accidents involving vulnerable road users are far more likely. According to these critics, Autopilot’s highway bias, where it’s most often used, distorts the safety data.
However, we have yet to see a critical urban incident involving FSD across North America over the years and millions of miles, suggesting that the data is likely accurate and that Tesla is approaching the 10x-safer-than-humans standard they are aiming for.
The Road Ahead
The question is, will Tesla’s description of “safer” ever return? Quite possibly, yes, but only as FSD matures and becomes more accepted by both regulators and society.
Tesla has faced multiple lawsuits over Autopilot over the years, and several in which it had to pay significant damages. This change is likely being pushed from within Tesla’s legal department. However, we would love for Tesla to stand up for what it advertises, what it believes in, and what their data shows.
