Tesla’s Own AI Trainers Don’t Trust ‘Full Self-Driving’ or Its Safety Stats, Reuters Finds

Why It Matters

The misrepresented safety metrics and internal distrust erode confidence in Tesla’s autonomous‑driving narrative, exposing the company to heightened regulatory risk and potential investor fallout.

Key Takeaways

• Reuters found Tesla’s FSD safety stats overstated threefold due to flawed comparison.
• Former data labelers say they wouldn’t trust FSD to drive them.
• Tesla mapped robotaxi zones, contradicting Musk’s claim of no local mapping.
• Age gap between Tesla fleet and US fleet skews safety metrics.
• NHTSA probes and $243 M verdict raise regulatory pressure on Tesla.

Pulse Analysis

Tesla’s claim that Full Self‑Driving is up to ten times safer than human drivers has long been contested, but Reuters’ latest investigation quantifies the exaggeration. By juxtaposing airbag‑deployment crashes in Tesla vehicles with a federal dataset that includes far less severe tow‑truck incidents, the company’s safety ratio inflates by a factor of three. A correct apples‑to‑apples comparison, as performed by University of Michigan researcher Marco Benedetti, still shows a modest three‑fold advantage, but even that is compromised by the Tesla fleet’s average age of just over four years versus the national average of nearly thirteen. This methodological slip not only misleads consumers but also raises questions about the rigor of Tesla’s internal analytics.

Beyond numbers, the human element paints a stark picture. Former data labelers—who scrutinize millions of video frames daily—report that they would not ride in a vehicle running FSD, citing repeated failures to yield to emergency vehicles, misjudging motorcyclist gaps, and reckless speed‑limit violations in “Mad Max” mode. The internal “trauma team” documented near‑misses with pedestrians and construction workers, underscoring a gap between public marketing and on‑ground performance. Moreover, Tesla’s robotaxi rollouts depend on labor‑intensive pre‑mapping of specific zones, directly contradicting Musk’s narrative that the system operates without detailed local maps. This approach limits scalability and suggests that widespread autonomous deployment remains a distant goal.

The fallout extends to regulators and investors. The National Highway Traffic Safety Administration currently pursues four active investigations into FSD and Autopilot, while a recent $243 million jury verdict over an Autopilot crash adds financial pressure. As safety claims lose credibility, Tesla faces heightened scrutiny that could affect its stock valuation and future autonomous‑vehicle partnerships. Stakeholders will be watching how the company adjusts its messaging, improves data transparency, and addresses the internal concerns of the very engineers tasked with teaching its AI to drive.