Tesla Achieves Musk’s Standard for ‘Safe Unsupervised’ Driving
Tesla has hit a significant milestone in its journey toward self-driving technology, with its fleet of vehicles utilizing the Full Self-Driving (Supervised) system surpassing 10 billion miles, as noted on the company’s safety page. This achievement marks a key point in Elon Musk’s vision for “safe unsupervised” driving.
However, Tesla owners won’t find their FSD (Supervised) vehicles magically upgraded to FSD (Unsupervised) overnight. The system remains a Level 2 setup, meaning it still requires a fully attentive human driver to monitor the road and be ready to take control at any moment.
Earlier this year, Musk mentioned that achieving around 10 billion miles of training data was essential for safe unsupervised self-driving. This led to speculation that reaching this milestone would trigger a switch to unsupervised driving for all customers.
Such a leap would be a risky move for Tesla, especially given the ongoing concerns about the company’s accountability for its fleet of over a million FSD-equipped vehicles. Unlike Waymo, which assumes liability for its technology and fleet, Tesla’s terms place that responsibility on the vehicle owner, primarily because FSD is classified as a Level 2 supervised system. This raises questions about who would be liable in the event of a crash if FSD were to operate unsupervised.
Tesla has yet to clarify this issue. Over the years, numerous incidents involving its partially autonomous features have occurred, resulting in fatalities. The company has managed to evade liability through settlements or court dismissals. On its website, Tesla emphasizes that FSD (Supervised) “requires active driver supervision and does not make the vehicle autonomous.”
A federal jury in Florida recently found Tesla partially liable for a fatal crash in 2019 involving its Autopilot system, ordering the company to pay $243 million to the victims’ families. Tesla has appealed this decision, but the appeal was denied.
Despite these challenges, the achievement of 10 billion miles driven in FSD (Supervised) is noteworthy. Tesla claims that its FSD-equipped vehicles average 5.5 million miles between major collisions, compared to 660,000 miles for the average driver in the U.S., suggesting that FSD is safer than human driving.
However, experts have raised concerns about Tesla’s safety reporting methods. Studies indicate that the company’s reports may overlook critical traffic statistics, such as the fact that crashes are more frequent on city and undivided roads than on highways, where Autopilot is predominantly used. Some researchers suspect that Tesla might be misrepresenting crash data to portray Autopilot and FSD as safer than they truly are.
While unsupervised driving remains out of reach for most Tesla customers, the company is expanding its use of unsupervised vehicles in its robotaxi fleet. After launching in Dallas and Houston with just a couple of vehicles, Tesla has increased its fleet size, with Dallas now operating five unsupervised robotaxis and Houston six, according to the Robotaxi Tracker. Austin, the initial launch city for the robotaxi service, currently has 29 supervised vehicles (with employees in the front passenger seat) and 22 unsupervised ones.
This development has many Tesla owners feeling that unsupervised driving is just around the corner. However, liability concerns are likely to keep this feature at bay for now. In a recent earnings call, Musk indicated that unsupervised driving would be available when “it is legal to do so,” and he predicted that unsupervised FSD in customer cars could arrive in the fourth quarter of the year.
Is this yet another milestone, or just another moving goalpost?