Because Waymo has absolutly no incentive to make themselves look better that they actually are, right? You are looking at marketing material 🙄
Obviously a fleet of robo taxis that have a very narrow operational area and so on will look somewhat good compared to general drivers that are driving on roads and during conditions no Waymo taxi has ever entered.
Sure, Waymo’s own numbers should be taken with a grain of salt; and obviously they will publish numbers that will make them look good. I’m with you there.
But at the same time, Waymo and other robotaxi companies are required by law to report any small accident (which makes sense, we want them to be safe)
Compared to human drivers, the IIHS estimates that about 1/2 of all property damaged and about 1/3 of injuries never get reported, for various reasons. So that skews the distribution in the other direction.
IMHO I see no reason to doubt Waymo’s numbers. Obviously, a machine, given enough training, will perform much better than a human. It never gets tired, it has a much larger training regime than humans, it won’t get distracted by a phone call, and it has better sensors. I’m a bit surprised that this point is highly contested.
I get that the economy side of it can be controversial (are we going to a world where no one owns cars? Is all the surplus going to alphabet? Etc), but safety seems like a clear win, to me.
Did you read the article linked? You are right that small accidents by humans are likely underreported, but any accidents where humans got hurt are usually also required to be reported to the police, and there human taxi drivers are 1.5 times less likely to be involved in such accidents. And IMHO I care a lot less about a small bumper hit where no one got hurt.
It is also false that machines necessarily have a lower accident rate. This would be true in a close system like an automated train, but in the messy roads where human drivers often can not sufficiently anticipate the often erratic driving of these automated systems, they are a risk factor that increases the danger.
Because Waymo has absolutly no incentive to make themselves look better that they actually are, right? You are looking at marketing material 🙄
Obviously a fleet of robo taxis that have a very narrow operational area and so on will look somewhat good compared to general drivers that are driving on roads and during conditions no Waymo taxi has ever entered.
A fairer comparison would be human driven taxis.
Edit: one such study shows Waymos are actually 1.5x more dangerous than taxis (including Ubers) driven by humans: https://www.techtimes.com/articles/322293/20260730/waymo-freeway-return-lands-amid-study-showing-higher-crash-rate-uber-drivers.htm
Sure, Waymo’s own numbers should be taken with a grain of salt; and obviously they will publish numbers that will make them look good. I’m with you there.
But at the same time, Waymo and other robotaxi companies are required by law to report any small accident (which makes sense, we want them to be safe)
Compared to human drivers, the IIHS estimates that about 1/2 of all property damaged and about 1/3 of injuries never get reported, for various reasons. So that skews the distribution in the other direction.
IMHO I see no reason to doubt Waymo’s numbers. Obviously, a machine, given enough training, will perform much better than a human. It never gets tired, it has a much larger training regime than humans, it won’t get distracted by a phone call, and it has better sensors. I’m a bit surprised that this point is highly contested.
I get that the economy side of it can be controversial (are we going to a world where no one owns cars? Is all the surplus going to alphabet? Etc), but safety seems like a clear win, to me.
Did you read the article linked? You are right that small accidents by humans are likely underreported, but any accidents where humans got hurt are usually also required to be reported to the police, and there human taxi drivers are 1.5 times less likely to be involved in such accidents. And IMHO I care a lot less about a small bumper hit where no one got hurt.
It is also false that machines necessarily have a lower accident rate. This would be true in a close system like an automated train, but in the messy roads where human drivers often can not sufficiently anticipate the often erratic driving of these automated systems, they are a risk factor that increases the danger.