Catch up on stories from the past week (and beyond) at the Slashdot story archive

 


Forgot your password?
Close
typodupeerror
×
Transportation

Waymo's Driverless Cars Are Apparently an Insurance Company's Dream (engadget.com) 146

Posted by BeauHD from the would-you-look-at-that dept.
A study by reinsurer Swiss Re found that Waymo's autonomous vehicles have demonstrated significantly fewer property damage and bodily injury claims compared to human-driven cars, with reductions of up to 92% in some metrics. Engadget reports: Swiss Re analyzed liability claims from collisions covering 25.3 million miles driven by Waymo's autonomous cars. The study also compared Waymo's liability claims to human driver baselines based on data from over 500,000 claims and over 200 billion driving miles. The results found that Waymo Driver "demonstrated better safety performance when compared to human-driver vehicles." The study found cars operated by Alphabet's Waymo Driver resulted in 88 percent fewer property damage claims and 92 percent fewer bodily injury claims.

Swiss Re also invented a new metric to compare Waymo Driver against only newer vehicles with advanced safety tech, like driver assistance, automated emergency braking and blind spot warning systems, instead of against the whole corpus of those 200 billion driving miles. In this comparison, Waymo still came out ahead with an 86 percent reduction in property damage claims and a 90 percent reduction on bodily damage claims.

Waymo's Driverless Cars Are Apparently an Insurance Company's Dream More | Reply

Waymo's Driverless Cars Are Apparently an Insurance Company's Dream

Comments Filter:

  • Hey! Rates will go down! (Score:5, Insightful)

    by buss_error ( 142273 ) on Thursday December 19, 2024 @07:54PM (#65027175) Homepage Journal

    with reductions of up to 92% in some metrics.

    Well, with this comes great news! The rates of coverage will only go up by 50%!

    • So the big question is if waymo vehicles travel safer routes than general driving. At least this is broken down by zip code . But one could probably come up with a different measure comparing accidents on similar paths. The data may not be dense enough for that though

      • Re:Zip code level data (Score:4, Interesting)

        by Frank Burly ( 4247955 ) on Thursday December 19, 2024 @08:36PM (#65027285)
        On my normal drives I see them so often that they either could provide the basis of an alibi or an invasion of privacy lawsuit. So, while I don't know where they're going, I know they are using my way of getting there.

      • Re: (Score:3, Interesting)

        by swillden ( 191260 )

        So the big question is if waymo vehicles travel safer routes than general driving.

        I think they actually use more dangerous routes, because they stay on surface streets. I read just a month or two ago that they have finally begun using freeways in Phoenix. I don't believer the use freeways anywhere else they operate.

        • Re: (Score:2)

          by rossdee ( 243626 )

          "I think they actually use more dangerous routes, because they stay on surface streets. "

          As opposed to the underwater streets? Or underground streets?

          • "I think they actually use more dangerous routes, because they stay on surface streets. "

            As opposed to the underwater streets? Or underground streets?

            As opposed to limited-access highways which don't have intersections in the normal sense of the word, and which are often grade-separated from streets which do have intersections and which might have pedestrians on them.

          • "I think they actually use more dangerous routes, because they stay on surface streets. "

            As opposed to the underwater streets? Or underground streets?

            Have you really never heard the phrase "surface streets" before? Perhaps you're not from the US and that phrase is an Americanism?

        • Do they get much snow in SF, LA, Phoenix, or Austin?

          • Do they get much snow in SF, LA, Phoenix, or Austin?

            I don't see how that's relevant. The human driver data was from the same regions, so it's a comparison of like to like.

      • I'm certain that's a factor. They're severely limiting the range within which the cars can operate.

        I want to take trips from my home with no cell signal to a remote beach also with no signal, going through a route with at least 20 miles of no signal. And there is of course no EV charger available at destination or en route . 2 slow 32 amp ones at my home. Only way I know yo make this trip is to drive myself. Uber won't work since I can't summon them on the way back due to no signal. I also can't drive in th

      • Re: (Score:2, Interesting)

        by SoftwareArtist ( 1472499 )

        It doesn't say their cars get in fewer accidents. It just says they result in fewer insurance claims. That's not the same.

        Just speculating here, but it wouldn't surprise me if they mostly self insure. They're big enough to pay for minor accidents out of pocket and only insure against major losses. If you can afford to do that, it's usually less expensive in the long run.

        We should be able to find out. At least in California, they're required to report every time one of their cars is in an accident. Pro

        • Re: (Score:3, Informative)

          by Sique ( 173459 )
          That's why the analysis comes from Swiss Re, which is - tada! - a re-insurer, an insurer for insurers to cover extremely large payouts. Swiss Re thus knows about all the large sums of money they had to pay for larger damage, which mostly means car crashes with heavy injuries or deaths. Swiss Re has no involvements into fender benders.

      • Or it might just be they are kitted out with so much camera coverage and other data collection that they can almost always prove the other party is at fault and are not liable for damages.

        This means they could actually be in more actual accidents but just almost always never have to pay out.

      • We already know the answer to this - the miles covered by Waymo vehicles are indeed safer miles than the general driving public.

    • Re: (Score:2)

      by kmoser ( 1469707 )
      And the rate of denied claims will only increase by a small amount, too!
    • or 100% in the USA!

  • Odd way to rate safety (Score:3)

    by votsalo ( 5723036 ) on Thursday December 19, 2024 @08:01PM (#65027185)
    Instead of saying Waymo car rides are safer, they are saying the insurance companies like them? Only insurance companies care about having fewer traffic deaths, injuries, and property damage? And why would insurance companies care anyway? Riskier tasks can command higher insurance premiums and profits.

    • Re:Odd way to rate safety (Score:4, Insightful)

      by Firethorn ( 177587 ) on Thursday December 19, 2024 @08:20PM (#65027237) Homepage Journal

      Higher premiums, maybe, but those profits are more risky, as there's a greater chance of screwing it up and losing money because of extra claims.

      They want to know the actual risk so they can compete with other insurance companies, charging enough to cover the risk, but still hopefully undercutting the competition.

    • Certainty not Safety, Average not Enough (Score:5, Interesting)

      by Roger W Moore ( 538166 ) on Friday December 20, 2024 @12:04AM (#65027521) Journal

      Instead of saying Waymo car rides are safer, they are saying the insurance companies like them?

      Insurance companies will like them regardless of their safety rating because they are a known quantity. There are many of them all with identical algorithms and so the rate and severity of accidents can be measured to far higher degree of precision than with a human driver since we are all different. In addition, unlike a human driver is not going to start driving bad because it had a major upset or because someone had to work late and drive home tired etc. This lets them know with a higher degree of certainty how much it will cost them to insure them and hence how much profit they will make.

      If you want to measure safety in a way that will persuade people that driverless cars are safe then the measure you want is what fraction of human drivers are safer than a Waymo. Waymo may result in ~90% fewer claims than the average of human drivers but if say (and I'm inventing numbers here to illustrate clearly why this matters) 90% of all human-driver claims came from the worse 10% of human drivers then that would make 90% of human drivers safer than a Waymo. While that example is probably too extreme the distribution of accidents amongst human drivers matters, the average is not enough information to know that driverless cars are safer than most human drivers.

      • This is true, and itâ(TM)s related to what I would guess are the biggest reasons for the difference in performance: (1) Computers donâ(TM)t drink and (2) They never look away from where theyâ(TM)re going (to text or change the radio or look at the weather or whatever). But mostly the not drinking part.

      • If you want to measure safety in a way that will persuade people that driverless cars are safe then the measure you want is what fraction of human drivers are safer than a Waymo.

        Two problems, first people aren't going to be convinced by statistics, second the majority of people think they are a better driver than average so it really only applies to the other guys. Cars are pretty dangerous compared to a lot of other things considered dangerous, and people's views are already very heavily skewed in order to

        • Re: (Score:2)

          by Sique ( 173459 )
          First, people are convinced by lower rates, and second, just because you believe to be a better driver, you don't necessarily want to drive. Waymo & Co. will convince people by convenience and price. If Waymos are cheaper than Ubers, people will call a Waymo. And if Waymos are cheaper than the running cost of owning a car, people will not buy a new car, but use Waymo instead.

        • Re:Certainty not Safety, Average not Enough (Score:5, Interesting)

          by Roger W Moore ( 538166 ) on Friday December 20, 2024 @08:06AM (#65027977) Journal

          Two problems, first people aren't going to be convinced by statistics, second the majority of people think they are a better driver than average

          Only some people are not convinvced by statistics and for the rest it depends a lot on what those statistics are. Yes, most people think they are better than average but if the statistics only show that a Waymo is only better than half of human drivers then that would definitely not be enough to convince me. By the time it is getting up to the level where it is safer than 95% of human drivers it becomes a lot more convincing: I might think that I'm better than average but I'm not sure that I am in the top 5%.

          It;s also worth pointing out that if you measure driver safety by the number of accidents then it is entirely possible for a majority of people to actually be a better driver than the average. In the hypothetical extreme example I gave where 10% of drivers caused 90% of accidents, more than 90% of drvers would have fewer accidents than the average rate thus making them "better than average" by this definition. This highlights the other problem with statistics: not many people really understand them and even those of us who do can easily get confused by them sometimes.

      • Re: (Score:2)

        by q_e_t ( 5104099 )
        At the moment, I'm be unconvinced that they like them. They have decades of data from humans driving cars, and that level of unpredictability on an individual scale can be calculated pretty well in aggregate. With self-driving cars, they don't have that information to go on, so there is a great deal of potential uncertainty should there be, perhaps, a bug in the software or sensors that emerges resulting in a big surge in accidents. It's low risk, high consequence, and insurance companies normally charge hi
    • Any company "liking" or "loving" or thinking anything is a "dream" is not mentioned in the engadget story, let alone the study in links to.

      Please remember at all times: headlines of almost all news outlets are clickbait bullshit. They normally suggest something more salacious than what the facts in the story actually are. Read the comments on any comment board, and 95% are replying to what the headline implies, which may or may not be what the facts of the story are about.

    • Re: (Score:2)

      by taustin ( 171655 )

      Riskier tasks can command higher insurance premiums and profits.

      Where on Earth did you get the idea that premiums are based on risk, or anything other than "whatever we can get away with charging"?

    • The specific data point here is that insurance claims are significantly lower with the Waymo cars than human drivers. There is a correlation between safety and insurance claims but it's not a coefficient of 1. If a vehicle has less fender benders but an equal number of fatal accidents, the insurance payouts would be lower with no safety improvement. Admittedly, that's unlikely, because accidents involving severe injuries tend to be the most expensive so one can infer that such a large reduction in payout

  • It's hard to get into accidents when they randomly decide to stop somewhere and not move.

  • Comparing apples with oranges (Score:5, Insightful)

    by VeryFluffyBunny ( 5037285 ) on Thursday December 19, 2024 @08:32PM (#65027273)
    This is disingenuous because they're not comparing like with like. They're comparing taxis with all cars. They're comparing routes where Waymo cars drive with all routes driven by all cars. If you want a like with like comparison, compare Waymo cars with taxis driven along the same routes. Let's see how the stats add up under those conditions.

    This is basic stats 101 & journalists really shouldn't be falling for this shit. It's embarrassing just to watch.
    • It's fine if all they want to do is calculate insurance fees. The headline is maybe not supported by the paper, but that's normal scientific reporting.
      • It's not fine even for that. Actuarial science 101 requires understanding probability theory and statistics. You can't do proper statistics with highly imbalanced data.

    • There is another dimension to this (Score:5, Insightful)

      by rapjr ( 732628 ) on Friday December 20, 2024 @03:15AM (#65027715)
      The number of accidents may be lower, but people may be less accepting of certain kinds of causes of accidents. A drunk driver killing your child is bad and the drunk driver goes to prison. An algorithm deciding a person wearing a patterned shirt is invisible and hence it drives through them is not something people will accept as understandable, and how can they hold the algorithm to account? Can they prosecute the algorithm so it never drives again? There is a human dimension to this that goes beyond the statistics. Until self driving vehicle researchers realize this, they are going to find it hard to gain acceptance for self driving vehicles. It amounts to allowing random murders because it's better than murders that have a cause. Technically true, but emotionally unacceptable.

    • Not as disingenuous as you think. Given the vast majority of accidents happen in the city and Waymo overwhelmingly drives in the city compared to all cars which drive on highways (with far safer stats) your complaint actually works in Waymo's favour.

      Now if you were talking about Tesla I'd agree with you, since their automated systems are overwhelmingly used on highways.

      • Re: (Score:2)

        by q_e_t ( 5104099 )
        I did a quick calculation for the UK, and roads (in general) are 6 times as dangerous as motorways (highways) per mile travelled in terms of fatalities. Whilst not the USA, I expect similar trends apply.
      • But that's still not comparing like with like. What are the collision stats for taxi drivers on the same routes as Waymo cars? That'd be comparing like with like.
    • Previous studies have found that real taxis and ride-hailing services have fewer accidents than general drivers in the same cities by about 30%-50%:

      http://www.schallerconsult.com... [schallerconsult.com]

      https://www.courtlaw.com/blog/... [courtlaw.com]

      Whereas this study found that Waymo has about 90% fewer accidents than the general population in the same zip code. It also notes that the roads Waymo doesn't go on have fewer accidents per mile, not more.

      So, comparing only to taxis closes some of the gap, but not all.

    • Because it's going to religiously follow every single law that it can. It'll drive the exact speed limit at all times where is when I get in an Uber I can tell you that guy isn't going the speed limit he's flowing with traffic like everybody else. And if there isn't enough traffic he's probably going to go a bit over the speed limit like normal people do because we intentionally set the speed limit lower than it really probably needs to be for perfect safety with the assumption that people are going to spee

  • Of course they're safer (Score:3)

    by quonset ( 4839537 ) on Thursday December 19, 2024 @09:28PM (#65027367)

    They don't pass one another or other vehicles [imgur.com].

  • Oh, sure, they're happy now (Score:5, Insightful)

    by RogueWarrior65 ( 678876 ) on Thursday December 19, 2024 @09:29PM (#65027371)

    But eventually, there will be no need for car insurance.

    • You say that like it is a bad thing. There may be very little need for police too, as I have observed that what, 80% of their time is car related.

    • Re: (Score:3)

      by PPH ( 736903 )

      'Insurance' will be provided with the car by the manufacturer. I'll be damned if I'm going to pay for the liability 100% created by someone else's software.

    • Re: (Score:3)

      by gweihir ( 88907 )

      There still will be. It will be a lot less profitable, but insurances are used to business changing.

  • so fewer injuries. also no worker's comp insurance necessary.

  • Has been predicted for a long time. Human drivers are, on average, really terrible. Eventually, this will lead to human drivers being the exception.

    • I've been saying (in print) for about 30 years now that the way they'll eventually get humans out of the driving seat is by charging them about 100k per year for insurance if they insist on taking the wheel.

      • Re: (Score:2)

        by gweihir ( 88907 )

        Indeed. The plans for that have existed for a while. As soon as the tech is generally ready (say, another 10-20 years), they will be implemented progressively.

  • Do Waymo driverless cars support a 'cross country escape from the police followed by media helicopters' mode ?

  • ...if a hacker causes all Waymo cars to crash at once against the nearest car, I don't think that insurance companies would be that happy. The safety of the IT infrastructure beyond this technology should be considered, when calculating the insurance risk.

  • Anything that's a dream for an insurance company just means they are charging too much. As Swiss Re guessed.

    Now put Waymo tech in passenger cars and turn it on if you've been drinking instead of becoming a menace to other people. Voila! tempory Taxi. And maybe lower premiums, or maybe insurance would help pay for it even. Honestly Waymo could make a lot of money offering this, they take over your car if you cannot drive, and it would have a big impact on society. Put every teenager in a Waymo-equipped car a

  • The results found that Waymo Driver "demonstrated better safety performance when compared to human-driver vehicles." The study found cars operated by Alphabet's Waymo Driver resulted in 88 percent fewer property damage claims and 92 percent fewer bodily injury claims.

    Swiss Re also invented a new metric to compare Waymo Driver against only newer vehicles with advanced safety tech, like driver assistance, automated emergency braking and blind spot warning systems, instead of against the whole corpus of tho

  • An car insurers perfect customer is the one that pays their small premiums and never gets into an accident.

  • Self driving cars are inherently a subscription based service. As an individual, you cannot own one, only rent some of the hardware. If insurance "likes" this, it is because they are getting a cut out of extracting more money out of consumers. The end result will be always the same - it will cost you, the consumer, more to get from Point A to Point B.

  • There are more variables to consider than per-mile collision rates. Consider:

    1. Being held liable for trapping passengers in unsafe situations: https://www.youtube.com/watch?... [youtube.com]
    2. Traffic citations for being incapable of understanding abnormal situations: https://www.youtube.com/watch?... [youtube.com]
    3. Extremely centralized liabilities. When a person gets sued for negligence, they can major portions of their household worth. Hundreds of thousands of dollars. When a major corporate entity is held liable for negligence,

  • I wonder how much of this may be caused by Waymo cars being easily recognizable by human drivers and those dirvers, knowing there is no human in the Waymo, drive safer around them. For example, a human driver may not cut off a Waymo car, or jump out in front of them, because they don't trust the self driving car. In effect, by being recognized as a self driving car, the Waymo's produce a small bubble of safer driving conditions around them (kinda like student driver signs, if you see one of those you ty

  • Ah, but they need to factor in the amortization of the capital cost of the car. It likely is still cheaper to have a human car and pay high insurance, than to own a waymo car and pay low insurance.

Slashdot Top Deals

It is better to live rich than to die rich. -- Samuel Johnson

Close