The Silicon Valley dream of a driverless future crashed into a suburban reality this week. A federal investigation has been launched after a woman was killed when her Tesla, reportedly operating in ‘Full Self-Driving’ mode, failed to detect a stationary fire truck on a California highway. The woman, a 45-year-old mother of two, becomes a statistic in the growing ledger of ‘autonomous incidents’. But behind the numbers lies a more uncomfortable cultural truth: we are rushing to hand over control to machines that are still, in many ways, learning to see the world.
I found myself standing at a traffic light in London yesterday, watching a taxi driver negotiate a roundabout with that peculiarly British mix of aggression and deference. He saw a cyclist, a pedestrian, a dog. He made eye contact. He read intention. What Tesla’s computer reads, apparently, is a probability distribution. And probabilities, as we know, have outliers. The woman in the Tesla was the outlier.
The investigation will focus on the software, the sensors, the decision-making architecture. But the real story is the gap between the headline-grabbing ‘Level 5’ autonomy and the everyday reality of Level 2 assistance. Tesla owners are beta testers, often untrained for the task, lulled into a false sense of security by a brand that promises more than it delivers. The term ‘Full Self-Driving’ is a marketing masterpiece, but on the ground it is a misnomer. The safety regulators are right to ask questions. But so should we, as a society. What are we willing to trade for the convenience of a car that drives itself? A life. It seems. And that is a price too high for a future that is not yet ready.
