Tesla’s Full Self-Driving (FSD) system is supposed to be the future of driving. Elon Musk has promised it would soon deliver “safe unsupervised” rides. But guess what? The very people training that AI don’t trust it to drive themselves. Seven out of nine former Tesla data labelers say they wouldn’t get in a Tesla running FSD—not even for free.

The AI Trainers’ Shocking Lack of Trust

These workers see the system’s failures up close, every day. They review hours of footage showing Tesla cars speeding, missing stop signs, and making dangerous mistakes. They witness Teslas almost hitting kids and failing to spot pedestrians. They watch cars drive 30 miles per hour over the speed limit on public roads. Yet, their concerns about speeding were ignored. Engineers and managers prioritized rare, unusual scenarios over this routine but deadly issue.

One former data labeler said bluntly: “We have all seen it fail.” A former Tesla self-driving engineer told investigators, “Definitely don’t trust Elon on this.” Another refused to ride in a Tesla robotaxi “even if you f***ing paid me.” That’s a bold statement from insiders who build the system.

Speeding and Safety: The Overlooked Crisis

Speeding is a big deal. It’s a problem on every trip and affects safety everywhere. But Tesla’s internal teams treated it like a minor issue. Instead, they focused on “edge cases” — odd road layouts or rare lighting conditions. This choice is puzzling. Ignoring common safety flaws while chasing rare scenarios can make the system dangerous on everyday drives.

Independent researchers have also questioned Tesla’s safety claims. The company says FSD is ten times safer than human drivers. But that number comes from comparing Tesla’s relatively new fleet against the average American car, which is much older and less safe. They also count only crashes where airbags deployed, ignoring many other accidents. When these factors are corrected, Tesla’s safety advantage shrinks dramatically.

The Robotaxi Rollout: A Slow and Painful Reality

Tesla’s robotaxi dream is far from reality. The company launched unsupervised robotaxis in Austin, Texas, but only in tiny, carefully mapped zones. These zones took months of hard work and hundreds of employees to prepare. Tesla’s Utah-based labeling team grew to almost 300 people just to map these routes.

After a year, only about 50 robotaxis operate in Austin, many still with human safety drivers. The system needs constant supervision. Despite all the hype, Tesla’s robotaxis cannot roam freely or handle unexpected situations safely. The massive effort behind these limited zones contradicts Musk’s claim that Tesla doesn’t rely on “laborious local mapping.”

Meanwhile, recent crashes caught media attention. Teslas on FSD have driven into lakes, off bridges, and nearly hit trains. These are just the incidents reported publicly. The internal footage shows far more failures that never reach the news.

The Trust Gap and What It Means for You

Why should customers trust a system that its own builders won’t trust? The AI trainers see the raw data and know where the system fails. They don’t see glossy marketing or optimistic timelines. They see a product still struggling with basic safety.

Federal regulators have taken notice. Over two million Teslas have been recalled for autopilot issues. Investigations are ongoing into FSD’s handling of emergency vehicles, stop signs, and poor visibility conditions. Tesla’s aggressive marketing clashes with the messy reality on the road.

The autonomy race is not just about technology. It’s about trust, transparency, and safety. Tesla’s competitors like Waymo publish detailed safety studies and admit their limits. Tesla’s internal doubts highlight a bigger question: Is the rush to deploy autonomous driving outpacing safety and readiness?

Looking Ahead: The Road Tesla Must Travel

Tesla faces a crossroads. It must close the gap between promises and performance. That means fixing widespread issues like speeding, improving validation processes, and increasing transparency. Regulatory pressure may force stricter safety controls and clearer rules for deploying autonomous systems.

For consumers, the message is clear: don’t assume full autonomy just because a car has FSD. The technology is still evolving, and insiders warn it’s not ready to drive unsupervised. The AI trainers’ refusal to ride their own creation serves as a powerful warning.

Will Tesla rise to the challenge and build a truly safe self-driving car? Or will the hype crash against reality? The next chapter in autonomous driving is unfolding right now, and everyone’s watching.