In a damning indictment of Tesla’s Full Self-Driving technology, the very people responsible for training the AI that powers it refuse to ride in vehicles using the system. A Reuters investigation uncovered a troubling disconnect between Tesla’s public confidence in FSD and the private skepticism of those closest to its development.
What Happened
Reuters interviewed nine former Tesla data labelers and one ex-self-driving engineer about their trust in Full Self-Driving mode. The results were striking: seven of the nine data specialists said they would not willingly ride in a Tesla operating on FSD. One respondent was particularly blunt, stating they wouldn’t ride in a Tesla robotaxi “if you f**king paid me.” These individuals spent months labeling thousands of hours of driving footage—the foundational work that enabled Tesla’s neural network to recognize road hazards, traffic patterns, and driving scenarios. Yet despite their intimate knowledge of the training data and AI capabilities, they lack confidence in the final product’s safety.
The insiders expressed concerns about edge cases, unpredictable behavior in complex scenarios, and gaps in the system’s decision-making logic. Their hesitation reflects a reality that many AI safety researchers have warned about: training data quality doesn’t guarantee real-world performance, especially in safety-critical applications like autonomous driving.
Key Points
Tesla has aggressively marketed FSD as a transformative technology, with CEO Elon Musk repeatedly promising full autonomy is imminent. The company has positioned itself as years ahead of competitors and has convinced thousands of customers to pay $12,000 to $15,000 for the feature. However, FSD remains firmly in “beta” status, and multiple accidents have been attributed to its limitations.
The skepticism from data labelers suggests internal awareness of the system’s shortcomings. These workers intimately understand what the AI has and hasn’t learned from training data. Their unwillingness to trust their own creation speaks volumes about confidence gaps within Tesla’s autonomous driving division.
What This Means
This revelation creates a credibility problem for Tesla. When the engineers and data scientists building a safety-critical system won’t use it themselves, investors and regulators should take notice. It raises questions about whether FSD is being oversold to consumers and whether Tesla’s timeline for true autonomy is realistic.
For the broader autonomous vehicle industry, the story underscores the importance of independent safety validation and transparent testing protocols. As companies race to deploy self-driving technology, insider perspectives on system reliability deserve serious consideration—even when they contradict marketing messaging.
