Rov_Scam
8hr ago
Tesla's FSD isn't the game changer that Tesla fans think it is; it's basically glorified cruise control. Driving a car isn't so much a physically demanding exercise as it is a mentally demanding one, and eliminating the need to operate the steering wheel only offers about as much value as eliminating the need to work the pedals. Even that wasn't as big of a game changer as automatic transmission; while both would become all but universal around the same time, for decades people were more willing to pay thousands of dollars for optional AT than to pay hundred for optional cruise control. And the demise of the manual transmission (in the US at least) was due to lack of consumer interest, while cruise only became universal because the adoption of electronic throttles on all vehicles meant that adding it was as simple as adding a button. But I digress.
The point is that until the mental aspect of driving can be automated, any autonomous system will be of limited utility. You mention the high loss of life due to car accidents. It would certainly be a good thing if autonomous vehicles could reduce this, but the dangerous nature of cars in general makes it difficult to fully automate. We currently have a system where each driver is almost 100% responsible for accidents caused by their cars. There are situations like Firestone Exploding Tires™ where car manufacturers have been held responsible, but these are edge cases, not garden-variety accidents. Removing the mental aspect from driving means removing any liability from the driver and shifting it to the manufacturer. Manufacturers will not accept this.
Consider a typical case where A & B are involved in an auto accident, with each sustaining injuries. The way it currently works is that a police officer would visit the scene and conduct a preliminary investigation and may make a preliminary finding of fault. If he determines that B is responsible, B will enter a claim with A's insurance company, and the insurance company will make an offer to settle. It's highly unlikely that they would contest the preliminary fault determination at this stage, but they may balk if A claims damages for injuries that they don't think are justified. If A is unsatisfied with the insurance company's offer, he can hire an attorney and sue B. At this stage, the insurance company will hire a lawyer to represent B, and the parties will take depositions and look at medical records. The accident report may be used as a guide, but A's lawyer will try to develop evidence to establish B's liability independent of the police report. B's lawyer may contest liability if the facts lean in that direction, but it's more likely that he'll try to challenge the damages to secure a low settlement amount. A has to prove that B was negligent. This basically means that there's a reasonable amount of caution that all drivers are expected to take, and that B failed to take the appropriate amount of caution. This could be because he was driving too fast, playing with the radio, or any number of other things that we all agree that careful drivers don't do. A trial is possible, but in normal circumstances once discovery closes both sides will negotiate a number that everyone can agree on, which will be within the limits of B's insurance policy. If A's damages exceed B's coverage limits, it's likely that A will settle for the maximum amount covered, unless B is independently wealthy and has significant assets. It goes without saying that since B is at fault in this example, he doesn't sue anyone. If he has collision insurance he will get compensation for damage to the car, but he won't get anything for medical expenses or pain and suffering.
Now let's assume A and B get in the same accident except both are in fully autonomous vehicles, A driving a Tesla and B driving a Honda. Both parties will file suits against Tesla, Honda, and whatever dealerships or subsidiaries are involved. This is not a negligence case but a product liability case. Neither plaintiff will have to prove that any defendant violated a standard of care, only that the product was defective and was a substantial contributing factor to their injuries. This will be easy for them to do; if two cars collide it is obvious that one of them did something wrong, and since neither driver had any control over his vehicle by design, the problem must be a defective vehicle. In theory, each defendant could conduct a forensic investigation and make the argument that the other car was defective and responsible for the crash. Realistically, unless it's a high value case or it's obvious that one car did nothing wrong, this is unlikely to prove anything that justifies the cost, and both parties will settle. The plaintiffs' attorneys are going to expect to get full value in these cases since insurance limits are unlikely to be an issue and, even if they are, the companies have deep pockets.
Each year, there are over 6 million reportable accidents in the United States. The average insurance payout per accident is around $30,000. My legal practice focuses on product liability, and there is no system I'm aware of where a solvent corporate defendant streamlines liability claims the way insurance companies do, so each of these accidents is going to generate a lawsuit. Actually 2 lawsuits, since both parties will be making claims now. And the effect of splitting claims among multiple defendants and not having to worry about insurance limits means the average payout will be higher. And there are legal fees to contend with, too. So assume that $30,000 becomes $40,000, and tack on $10,000 in legal fees, then double it to account for both cars involved, and we get to a nice round $100,000 per accident. (I'm keeping things simple and ignoring the effect of multi-car pileups, single car accidents, injuries to passengers, and any number of other things). More than triple the current amount, the cost shifted from drivers to auto manufacturers.
Tesla has already been successfully sued in relation to their cars' autonomous capabilities, and in one case the plaintiff was awarded an eye-watering judgment. I don't want to talk too much about those cases though, because they aren't really relevant to what I'm saying here. Those cases proceeded on a theory that Tesla failed to adequately warn the user and breached various implied warranties. As such, Tesla was at least afforded the opportunity to argue that the drivers had been adequately warned of the system's limitations and that they were responsible for those accidents by failing to heed those warnings. One can argue over whether Tesla's marketing creates a misleading impression of the capabilities of their products. What is beyond argument is that, in the case of a true autonomous system, Tesla, or anyone else, will have to end the throat clearing and market it on the basis that one can treat it like a personal taxi and ride even if they're asleep, or drunk, or don't have a license. Regardless of technical abilities, you have to add in the additional cost of taking on an additional $600 billion of added liabilities if these things become universal. That's a huge cost for one industry to bear.
One obvious argument is that self-driving cars are safer than manual cars. Okay, but how much safer? Remember, they have to take on approximately triple of what is currently paid out in insurance and lawsuit settlements for accidents based in the US. Reduce that number by 50%, or 80%, and it's still huge. And it underestimates how safe cars actually are. The 6 million+ reportable accidents a year sounds like a lot. But that's in a country that drives over 3.3 trillion miles annually. It's one accident for more than a half a million miles driven. If you drive 15,000 miles per year every year between ages 18 and 85, you can expect to be involved in two reportable accidents in your lifetime. Any assumptions about the relative safety of Tesla's current offerings versus normal cars is based on a woefully inadequate sample size. I'm sure the thing is impressive, but how many miles have you driven it so far? 10,000? 20,000? I'm guessing fewer than even that, but even if you drove it 50,000 you'd still have to drive it more than 10 times as far before we'd expect even one accident.
The final question we have to ask is what accidents are we concerned about preventing? I ignored this for the sake of simplicity, but the statistics I gave are for reportable accidents, while the average insurance payout includes all accidents. I only used reportable numbers because these are more serious, but in a large percentage of these, the police are only involved because one of the cars is undriveable. I was in two accidents in my life when my car was totaled, but other than the hassle of having to get a new car and deal with insurance, I was completely fine. I don't want to suggest that we shouldn't worry about these kinds of accidents at all, but they're certainly less of a priority than ones that cause serious injury or death. There are obviously solutions to these problems; I don't want to suggest that the liability questions are completely intractable. After all, we're already paying for motor vehicle insurance, and the cost will simply be shifted to another entity. Pay Tesla or pay Nationwide, take your pick, it's all the same. The overall point isn't that companies can't do this, it's that they won't. None of them are confident enough in their products at present to make this kind of guarantee, and consumers used to doing things a certain way will balk at the high premium they have to pay to cover the insurance premiums of a third party, even if they've technically been doing so for every company they've ever bought a good or service from.
Not as long as you keep the lights off and don't cross swords.
More options
Context Copy link