Another Tesla driver is dead, and once again the company is pushing back on the suggestion that its Autopilot system bears any meaningful responsibility. The fatal crash in Texas has reignited a debate that the automotive and technology industries have successfully avoided resolving for nearly a decade: when a car can drive itself most of the time but requires human oversight all of the time, who is actually accountable when things go catastrophically wrong?
Tesla's response follows a familiar playbook. The company emphasizes that Autopilot is a driver-assistance feature, not a self-driving system, and that users agree to remain attentive with hands on the wheel. This framing is legally convenient and technically accurate. It is also, critics argue, fundamentally dishonest about how the technology is marketed, named, and actually used by the humans who purchase it.
The Naming Problem
Tesla's decision to brand its advanced driver-assistance system "Autopilot" — and later to offer a "Full Self-Driving" package — has long been criticized as dangerously misleading. The National Transportation Safety Board has repeatedly called out the gap between Tesla's marketing language and the systems' actual capabilities. When you call something Autopilot, you are invoking aviation's highly automated cockpit systems. When you sell something called Full Self-Driving, you are making an implicit promise about what the technology can do. That Tesla's terms of service disclaim these implications does not change how ordinary consumers interpret the branding.
The result is a peculiar form of responsibility arbitrage. Tesla captures the market value of perceived autonomy — the premium pricing, the brand cachet, the stock-price boost from self-driving hype — while contractually offloading the liability onto individual drivers who are told, in fine print, that they must remain vigilant every second.
The Regulatory Vacuum
The more troubling issue is not Tesla's marketing but the regulatory environment that has allowed this ambiguity to persist. The National Highway Traffic Safety Administration has opened numerous investigations into Autopilot-related crashes but has not established clear standards for what Level 2 autonomous systems must do, how they must be monitored, or how liability should be apportioned when they fail. State-level rules vary wildly. The federal government has largely deferred to industry self-regulation, treating autonomous vehicle development as an innovation priority rather than a public safety challenge.
This approach made some sense in the early 2010s, when the technology was nascent and regulators wanted to avoid stifling experimentation. It makes considerably less sense now, when millions of vehicles with semi-autonomous capabilities are on American roads and the death toll from Autopilot-involved crashes has reached several hundred over the past decade.
The Attention Paradox
The deeper problem may be architectural. Semi-autonomous systems like Autopilot create what human-factors researchers call the "automation paradox": the better a system works most of the time, the harder it becomes for humans to remain vigilant during the rare moments when intervention is required. Asking a driver to monitor a system that handles 99 percent of situations flawlessly is asking for a kind of sustained attention that human cognition is not designed to provide. The technology's success creates the conditions for its failure.
Tesla's driver-monitoring systems — which use cabin cameras to detect inattention — are an attempt to address this, but they are treating a symptom rather than the underlying design tension. The fundamental question remains: should we deploy systems that depend on human oversight but are engineered in ways that make effective human oversight psychologically improbable?
Our take
Tesla is not wrong that drivers bear legal responsibility for operating their vehicles. But the company has spent years cultivating expectations it now disclaims, and regulators have spent those same years declining to establish the rules that would force clarity. The result is a system where Tesla gets the upside of autonomy hype while individual drivers — and their families — absorb the downside of fatal edge cases. This is not a technology problem. It is a governance failure, and another Texas death will not be the last until someone decides that the question of liability in semi-autonomous driving deserves an actual answer.
