Summarize this content to 2000 words in 6 paragraphs in Arabic Unlock the Editor’s Digest for freeRoula Khalaf, Editor of the FT, selects her favourite stories in this weekly newsletter.The road towards self-driving cars is paved with promises to save humans from themselves (and each other). Waymo, Google’s self-driving project, says that “the status quo of road safety is unacceptable” and that autonomous driving “can save lives”. Elon Musk, never one to be outdone, said in October when unveiling Tesla’s Cybercab that autonomous cars would become “10 times safer than a human” and “save lives — like, a lot of lives.”It’s a worthy goal, and one that sounds achievable. After all, humans are terrible drivers, aren’t they? In 2022, the latest year for which there is detailed data, 42,514 people were killed in motor vehicle traffic crashes on US roads, according to the National Highway Traffic Safety Administration. Of those fatalities, 29 per cent were killed in speeding-related crashes, and 32 per cent were classed as “alcohol-impaired driving fatalities”. Fully self-driving cars, in contrast, don’t drink and drive. They don’t speed because they’re impatient or running late. They don’t get sleepy. They don’t get distracted by their phones. On top of that, they have 360-degree vision. “You have to have eyes in the back of your head,” I remember a truck driver telling me once. Autonomous vehicles actually do.But for the companies which make the vehicles, there are two problems with their promise to save lives. The first is that human drivers have set the bar higher than you might think. Once you consider how many miles we drive, fatalities are actually very rare. The fatality rate per 100mn vehicle miles travelled in the US was just 1.33 in 2022. It has been close to 1 for about a decade. The latest data from Waymo shows its vehicles had only driven 33mn miles without a human driver through to the end of September last year. This means it simply hasn’t driven anywhere near enough miles yet to make a statistical comparison with human drivers when it comes to fatality rates.That said, Waymo’s record on less-serious accidents looks encouraging. The company’s data analysis suggests its vehicles have had fewer injury-causing crashes and police-reported crashes so far compared with humans driving in the same cities. Waymo told me it was “already significantly reducing the number of serious crashes compared to human drivers in places where we operate”. But Phil Koopman, an autonomous vehicle safety expert, cautions against complacency. “We know that computers fail differently than people,” he told me. “Human drivers fail individually, but [with] computers, every car has the same driver, so something can happen that causes them all to fail.” In addition, he says, computers might not get drunk, but nor do they have “common sense.”That brings us to the second problem. Even if self-driving vehicle companies do amass enough data to show their technology is saving lives, there are still widely-publicised examples every now and then of self-driving cars making bizarre mistakes that most humans simply wouldn’t make. There was the Waymo taxi which drove round and round in circles in a car park with a confused passenger inside, for example. Or the time that not one but two Waymos bumped into the back of the same pick-up truck which was being towed down the road at a slightly unusual angle, confusing the software. Most seriously, there was the time a pedestrian was thrown by another vehicle into the path of a General Motors Cruise self-driving car, which then dragged her for 20ft underneath it. In that case, a subsequent review commissioned by Cruise found that “an alert and attentive human driver would be aware that an impact of some sort had occurred and would not have continued driving without further investigating the situation”.It might be unfair, says Koopman, but it is human nature that stories like these stick in the mind more than statistics. If your selling point is “our cars are safer drivers than humans”, then every time one of your cars does something weird that a human wouldn’t do, your case with the public is eroded. Conversely, the accidents that didn’t happen thanks to your technology rarely make the news. That might be why annual surveys by the American Automobile Association show that people are becoming more fearful of the technology over time. The share of US drivers who say they trust self-driving vehicles has dropped from 14 per cent in 2021 to 9 per cent in 2024, while the share who say they are “afraid” has risen from 54 to 66 per cent. Self-driving car companies might say they want to save lives, but it’s going to be harder than they think to persuade people to let them give it a try.sarah.oconnor@ft.com
