Three weeks after one of Waymo's autonomous vehicles drove itself into a telephone pole for no apparent reason whatsoever (prompting 672 cars recalled), Waymo has released a statement saying its AVs are still 200 to 350% better at avoiding crashes than humans.
Waymo, a subsidiary of Alphabet Inc. (the parent company of Google), was formed in 2016 as a spin-off of Google's self-driving car division. In 2020, Waymo became the first company to offer service to the public as a fully self-driving taxi service.
The autonomous vehicle company has been plagued with negative headlines throughout its short history. By 2021, the National Highway Traffic Safety Administration (NHTSA) had recorded 150 crashes involving Waymo vehicles. In December of 2023, two Waymo robotaxis crashed into the same truck being towed on a highway only minutes apart. In February of this year, one of its cars struck a bicyclist in San Francisco. The latter two incidents prompted a recall of 444 of the company's cars.
There's been plenty of public backlash with the service, and numerous incidents of people interfering with and vandalizing the cars in protest under the guise that Waymo AV's are unsafe. Cars have been coned off to block their path, people have covered up the car's sensors, and there have even been reports of people setting them on fire.
Last month the NHTSA launched an investigation into Waymo, focusing on 31 incidents including the robo-taxis smashing into closed gates, driving on the wrong side of the road, and being the subject of car fires.
This month, Waymo posted to X (formerly Twitter) that its vehicles, compared to humans driving the same 14.8 million-mile (23.8 million-km) distance, had 30 fewer injury-causing crashes and 32 fewer police-reported crashes. The company claimed that these figures made their vehicles 3.5x and 2x safer, respectively, than vehicles driven by humans.
The breakdown is the sum of 3.83 million San Francisco miles with 17 fewer injury crashes and 12 fewer police-reported crashes, along with 10.92 million Phoenix miles with 13 and 20 fewer, respectively. Waymo currently operates its driverless cars in both cities, and expects to offer service in Los Angeles and Austin, Texas soon.
Having personally lived and driven extensively in San Francisco, I can attest that the streets of SF resemble that of an amateur destruction derby, leaving no car unscathed.
Currently in the USA, approximately 41k people die in auto accidents each year plus hundreds of thousands more suffer wounds that can disable them for life.
Even if AI driven cars caused 10k deaths to start with, that would be a 300% improvement over what we live with now!
THIS is how the benefit of AI driven cars should be presented!
> [...] what the US needs is more public transportation anyway.
And by having those public/mass transit vehicles automated we'll ensure there's drivers to handle it.
And where I live (in the center of mainland USA), there's very little public transportation for the majority of my state. Even the largest cities have public transportation in only a limited area of the city. If I wanted to take the bus (the only public transportation option I'm aware of) from my house (suburbs) to a job in the heart of the city, I'd first have to drive about 6 miles to the nearest bus stop, get on the one (ONE!!) morning bus from that location, ride for another hour while it makes many stops, finally arriving at the bus stop that's closest to my office (and even then it would probably be 6+ blocks) to walk. And then I repeat the process in reverse at night because there's only one stop at the bus station where I left my car at in the morning.
I don't know what the economics and efficiencies are, but it seems if we could take a 50-person bus and break it into five automated vehicles that could drive a route and adjust as needs demand (send a couple to my subdivision and pick us up like a parent taking a carload of kids to school), then transfer to that big bus in the morning. And maybe in the evening I have to work late, the same small vehicle could pick me and a couple others up downtown and make individual stops on a more optimized route.
Hey, I can dream, can't I?
It's not being a Luddite to say autonomous driving should be perfected first BEFORE it was ever put into public use. Not to mention the stats being bandied around are dubious at best, because we're not allowed to see the actual mileage those cars drove without drivers. We're just expected to take the Venture Capitalists at their word, and they never ever lie... right?
But like all tech advances over the past 25 years, we're going to run pell mell towards a future no one asked for, few want, and which doesn't resolve all the underlying problems plaguing our society. Build robust mass transit systems, streets that are walkable with shops that are accessible, and scale our world to people not automobiles. You want autonomous? Bring back trolleys and give them computer drivers. We should be moving AWAY from the automobile, not running towards an even more dystopic version of them.