Waymo Says Its Self-Driving Cars Reduced Injury-Causing Accidents by 82% (electrek.co)
- Reference: 0185873558
- News link: https://tech.slashdot.org/story/26/09/27/0134224/waymo-says-its-self-driving-cars-reduced-injury-causing-accidents-by-82
- Source link: https://electrek.co/2026/09/24/waymo-says-it-has-stopped-841-injuries-in-271-million-autonomous-miles/
[2] Electrek reports :
> We've seen various Waymo crash data before, with Waymo [3]claiming crash reductions . That's all well and good when the company says it, but we've also seen independent data confirming similar (though lower) [4]crash reduction numbers ...
>
> Waymo has enough miles that it's ready to start quoting how many injuries it has prevented, and the number is pretty high. Its newest crash data states that it had operated a total of 271 million driverless miles through June of this year, which is 50 million more miles added in the 3 months since its end-of-March update. Over those miles, Waymo says there was an 82% reduction in crashes that caused injury, and a 95% reduction in crashes that cause "serious injury or worse" [compared to human drivers].
Waymo also says that compared to human drivers it's reduced injury-causing crashes involving pedestrians by 93%, cyclists by 86%, and motorcyclists by 82%.
Waymo's analysis comes from San Francisco, Los Angeles, Austin, Atlanta, and Phoenix, and its [5]blog post includes video showing some near-misses where it says its automated system prevented an injury-causing collision.
[1] https://waymo.com/blog/shorts/safetydata-september26/
[2] https://electrek.co/2026/09/24/waymo-says-it-has-stopped-841-injuries-in-271-million-autonomous-miles/
[3] https://electrek.co/2023/12/20/waymo-says-its-self-driving-cars-are-safer-than-human-drivers-by-up-to-10x/
[4] https://electrek.co/2026/07/25/waymo-is-2-3-safer-than-a-human-driver-says-iihs-with-some-caveats/
[5] https://waymo.com/blog/shorts/safetydata-september26/
Pedestrian or vehicle accidents? (Score:2)
Just wondering, since related topic is Tesla FSD. EU is right now debating on whether to allow FSD across all member states. Some member states have permitted it already, but the big decision is still to be made. There's an interesting report from Belgium where Tesla FSD decided to disobey traffic laws around pedestrian/cyclist zones (going 50 kph in 20/30 kph zones, overtaking cyclists in bicycle lanes, etc).
[1]https://www.johanna.be/images/... [johanna.be]
Results start on page 12, but summary include the nice snippet
"Per
[1] https://www.johanna.be/images/articleimages/FSD/report_FSD30.pdf
Re:Pedestrian or vehicle accidents? (Score:5, Interesting)
[ Disclaimer: I'm a Tesla driver, since 2018, and have FSD on my car ]
Tesla's FSD and Waymo's approach are wildly different, enough so that you can't really compare them. The biggest two differences are: 1) Musk's cheap, so Teslas only use cameras to figure out what's going on around them; Waymos use cameras, ultrasonic sensors, radar, and lidar; 2) Waymos are only deployed on specific roads and streets that have been very heavily surveyed and Waymos have been trained on.
I trust Tesla's FSD about as far as I can throw it, and while it's been helpful and prevented an accident or two for me when I've been distracted, it's basically a glorified lane assist and cruise control. I'd not hesitate to fall asleep in the back of a driverless Waymo.
Re: (Score:2)
dafuq's wrong with you, friendo?
Yeah right, more like 100% (Score:2)
I believe it is 100% because they don't have a driver that can get hurt!
comparable (Score:5, Insightful)
Does anyone know what numbers they are comparing against? For this statement to be accurate they would have to measure humans actually driving in the exact places they drive and in the weather they drive. Different weather is not comparible, different roads are not comparable. Even if the human is in an older car that is not comparable.
Re: comparable (Score:2)
ok but are these all roads that waymo drives on with modern vehicles? you didn't really answer the question. they need to omit crashes in driveways, parking lots as well unless they navigate there.
Re: (Score:2)
Waymo doesn't count an entire class of accidents, a class far more common to occur at slower speeds. Of course, it's then important to know what Waymo's mix of slow speed vs high speed driving is compared to the human data. This is lying with statistics.
Re: (Score:2)
"Simple, the amount of crashes caused by a human or a driverless car over an x amount of miles in a city environment. "
This is certainly false, both because of the bogus assumption what the miles are and the lack of information on what's a "crash" and what "caused by" means.
"People really are bad at driving generally..."
This is wildly false, it's an Elon Musk-level lie.
"almost all accidents are caused by a human driver not paying attention while driving, not due to some technical failure of the car itself"
S
Re: comparable (Score:2)
They use various vehicle miles travelled bench marks published by state police in the regions the serve: [1]https://waymo.com/safety/impac... [waymo.com]
[1] https://waymo.com/safety/impact/#methodology
Re:comparable (Score:4, Interesting)
> Different weather is not comparible
Across 271million miles it's safe to assume the weather is similar. But even if the answer is something stupid like Waymos don't drive in the rain (which they do, they only stop during really severe storms), ultimately that's still injury avoidance as it just shows poor judgement of the people who drive under dangerous conditions.
Re: (Score:3)
Does it though? Waymo can refuse a ride to the emergency room because of a shower, you would not. Sociopathy is not a virtue...unless you're Elon Musk.
It's never a question whether self-driving should ever even be considered on roads designed only for human drivers. I think it should NOT be. A mile high stack of band-aids piled on top of a software stack that no one understands and that contains a limitless number of bugs that no one will ever fix is not my idea of a foundation to risk my life on, far b
Re: comparable (Score:2)
That people would take a Waymo to the emergency room because ambulances are so expensive is a separate problem
Re: (Score:2)
> Waymo can refuse a ride to the emergency room because of a shower, you would not.
I absolutely wouldn't ride to an emergency room. If I needed to go to an emergency room I'd call a professional taxi service by dialling 911.
The reality is people DO have poor judgement. 99.9% of people on the road are not there because their lives depend on them being on the road at that given time. Add to that "dumb people" statistic, i.e. those who try to drive their cars through flooded rivers and need to be rescued as a result, it all adds up to something awkward we don't want to admit: people are stup
Re: (Score:2)
You ludditeic anti-self driving people crack me up. First you guys were claiming a self driving car would never work anywhere. Now that the cars have proved themselves even in cities like San Francisco where the pedestrians and bicyclists are certifiably insane you're grasping for straws. Your pivoted claim is like "but, but, humans are better in driving in blizzards .. AND what about hurricane driving!1!@!!@!1one1! .. humans can beat them in that!". LOL.
Re: (Score:2)
Runner-up in the most horrifying neologism contest after "agentic": "ludditeic".
Re: (Score:2)
Seriously? Come on man, I thought it was brilliant. I thought of it a month ago and was waiting for a moment to use it without it being contrived or sounding forced.
Re: (Score:2)
It does harmonize well with "agentic".
Re: (Score:2)
I tried to look into this and found no information at all. Waymo doesn't report its mix of driven miles and how it compares to the human data. And what exactly qualifies to get reported? And Waymo can cause an accident without being involved and that would not be reported here, yet a human would be. Autonomous driving has incidents of substantial interference with traffic, there would certainly be incidents of Waymo causing damage to traffic that would not be reported.
It's easy to reduce traffic injuries
Another way Waymo reduces injury accidents (Score:2)
By clogging streets and preventing other cars from using the roads, Waymo helpfully ensures potential drivers remain uninjured!
[1]https://www.yahoo.com/news/us/... [yahoo.com]
[1] https://www.yahoo.com/news/us/articles/dozens-waymos-clog-presidio-july-142239118.html
Re: (Score:2)
I wondered why they don't randomise their routing a bit to prevent exactly this.
Re: (Score:2)
This is the most important thing for people to understand. Built into this data, carefully designed to mislead, is a misrepresentation as to what's important. Injuries can be reduced just my eliminating driving entirely but that doesn't make Waymo money. But Waymo ruining YOUR driving is perfectly OK. And they do.
I think it's hysterical that people think insurance companies will force drivers off the road. They will certainly raise prices as much as they can, but they will NOT destroy one of their most
I believe (Score:2)
I can easily believe that they have less accidents. Automatic systems do not get tired, distracted by fighting children, drunk, and can generally be assumed to obey the speed limit. The issue is all of the occasions when they face events that their system cannot handle, like inaccurate maps, bad weather, bad roads, etc. Of course, not driving under those conditions is an option, but people still want to get where they are going. I would like to see how they perform on the dirt country roads around here.
regarding driverless cars immobility (Score:1)
you know what would be an easy fix? equip waymos with a hand-held remote control unit that first responders can get to and drive the car. there are problems with this approach, of course, like waymo would probably have to remotely OK such a problem. still, seems like a pretty handy tool
Sleight-of-hand? (Score:3)
1st, comparisons against humans is an easy win. But Waymos aren't humans, they're machines. How happy would we be if refrigerators killed as many people as self-driving cars? Shouldn't we expect NO deaths from a machine? 2nd, being 'involved in' an accident doesn't tell us who's at fault. We need to know how many damage, injury and death accidents were the fault of Waymo. Stop comparing machines to humans isn't relevant.
Re: (Score:2)
Agreed on the sleight of hand but I have to say my kitchen fridge freezer has a "child-proof lock" to prevent children opening the doors unless they have the app. The manual says one of the reasons is to prevent children becoming trapped inside and "killed by the fridge". Though it's interesting that the US 1956 refgerator act was designed to prevent fridges from having lockable latches for the same reason (the child-proof lock on my fridge complies by intending to prevent children from opening the fridge i
Re: (Score:2)
> How happy would we be if refrigerators killed as many people as self-driving cars?
Well given Waymos have killed zero people, and fridge related deaths are hard to come by they have been attributed to fatal house fires and crushing incidents... I'm not sure you're making the point you think you're making.
82% (Score:2)
I guess the 18% is when they get hit from behind by a dude.
Yes, probably. And they likely can do even better. (Score:5, Interesting)
The main argument for self-driving cars is that human drivers are, on average, really bad. And that most of the bad human drivers think they are really good and hence add lack of care to bad skills.
I would also like to remind everybody that a few years ago, I think the CEO of Ford predicted that self-driving will be eventually enforced by the car insurers, because premiums for humans will become unsustainable. Looks too me like we are nicely on track.
Re: (Score:2)
I see no evidence of that. I see claims. Are Waymo's miles directly comparable to human miles? Who knows.
And human drivers are NOT, "on average, really bad".
Re: (Score:1)
They are in cities. Waymos don't drive highways all that much to my understanding.
And yes, humans are very good on average. But difference between (for example, number for illustration purposes only) 99.9% good and 99.99% good when it comes to millions of crashes every year is a meaningful number.
Re: (Score:2)
re: And human drivers are NOT, "on average, really bad".
Compared to what? It's meaningless without a relative reference. Compared to humans, their "average" (mean or mode?) is going to be "average" but that mean/mode may be significantly lower than something else, something like Waymo.
One thing humans do tend to do is overestimate how good they are. Average to poor drivers do tend to think they are better and numerous studies have demonstrated this. And if you want to see this in action, try keeping to the
Re: (Score:2)
There are many things humans are bad at.
Humans can't lift heavy objects
Humans can't drive
Humans can't go without eating
Humans can't go on repeating the same task without getting bored
Humans can't remember exact details consistently
Humans can't operate continuously without sleep
Humans can't handle extreme heat, cold, pressure, or radiation
Humans can't measure things with high consistency
Humans can't reproduce identical products or results
etc, etc
Fortunately or unfortunately if you start replaci
Re: (Score:2)
What would cause it to become unsustainable when it currently is? It will definitely be cheaper for a lot of people to get a car that is incapable of manual control, but insurance will also become cheaper for good drivers as a large chunk of what they pay is to cover for the bad drivers. I think future generations will largely avoid learn to drive unless they need to as a part of their job because a long commute isn't nearly as bad if a person doesn't need to operate a vehicle in traffic. But as self-drivin
Re: (Score:2)
> but insurance will also become cheaper for good drivers as a large chunk of what they pay is to cover for the bad drivers.
Bahahahahahaha, someone thinks something will go down in price. Oh you sweet summer child, come sit at the campfire, let me tell you a story about capitalism.