Former Tesla employees who worked on training and developing the company's autonomous driving systems have raised doubts about the safety of Full Self-Driving (FSD). According to a special report by Reuters, journalists interviewed nine former data labelers and one former Autopilot engineer, and their assessments were nearly unanimously skeptical.
Seven of the former labelers admitted they wouldn't ride in a Tesla with FSD engaged. One interviewee said employees saw FSD make mistakes, and called Elon Musk's claims about Tesla cars’ ability to drive safely unsupervised something that should absolutely not be believed. Another former worker was even more blunt, stating they wouldn't get into a Tesla Robotaxi even if paid.
Reuters reports that part of these workers' jobs involved analyzing large volumes of video recordings from FSD trips. Tesla used such material to train its software, teaching the cars not to repeat previous errors. However, former employees claim they regularly saw incidents of speeding and other problematic episodes. They added that some violations weren't always considered a priority by engineers and managers compared to rarer extreme scenarios.
These statements come amid ongoing debates about FSD's real capabilities and the future of Robotaxi. Tesla has long promoted the idea of fully autonomous transport as a key business direction, and Musk has repeatedly promised that the company's cars would be able to drive safely without a driver. But the testimony of people who worked with training data and system errors once again raises the question of how close Tesla actually is to that level of autonomy.
This is a particularly sensitive topic for the company, as trust in FSD is built not only on technical demonstrations but also on safety statistics that Tesla regularly cites in public statements. If even former development participants are unwilling to use the system in real trips, it reinforces critics' doubts and could complicate the perception of Robotaxi before its mass launch.
The main challenge for Tesla now is not just improving the algorithms, but convincing the public, regulators, and its own users that the technology is predictable. Autonomous driving requires not just impressive videos and bold promises, but proven reliability in everyday road situations, where the cost of a mistake can be too high.