What if we could make every car $5-10k more expensive instead?
</satire>
I once thought that using AI to make all cars self driving might be the key to making pedestrians and cyclists safe. Self driving cars are playing a game of Russian Roulette. The systems will get it wrong occasionally, with LIDAR or without. Not if, but when. Whether or not someone dies depends on the situation.
Tens of thousands of people die in car accidents per year, nationally, so a few dozen fatalities in exchange for faster self driving progress sounds like a good bargain.
If the car fleet kills half as many people as a human with a $1000/car system, and zero people with a $100,000 system, then we should immediately put the $1000 system on 100 times as many cars as we could put the $100,000 system on. (I picked those numbers because that is the order of magnitude range I have heard self driving car companies quote.)
The $100,000 system would only ever make sense if the fleet was already entirely self driving, and money for other life saving stuff like the environment and health care also hit suitable diminishing returns. Of course, by then, the cheap systems will have improved.
This argument holds for any non-negative dollar value you place on human life.
It is also independent of who owns the vehicles. Money the bus fleet spends in expensive self driving pulls money away from bus stop upgrades, pollution controls, etc, etc.
I think you're severely overestimating your case, not to mention the cost will probably be determined by a multitude of factors, one of the big ones being how advanced the self-driving tech is. Most of your points are just theoretical anecdotes.
You're not counting the RD cost that needs to be recouped.
You're overestimating the trust people will have with "oh didn't that once kill a person"-autopilot will have.
Wouldn't that make insurance cheaper for autonomous cars and more expensive for human-driven cars since humans would be more of a liability behind the wheel?
You are absolutely correct. That would almost certainly reduce deaths, and I would support taxing myself more for more money to go towards safer public transportation to reduce traffic fatalities.
I will, at the same time, support self-driving car research.
I think we could start putting self driving cars on the road today that would lower both the number of fatality's and accidents. We don't need more processing power, better sensors, or lower costs. What we need is slightly better software and the willingness to put it into production.
After-all the 'worst case' in a car is basically solved 99.9% of the time by staying in the correct lane, obeying stop lights / signs, speed limits, and simply hitting the breaks if your going to hit something. Sure, you could improve on that, but get that to work reliably and your already doing better than human drivers, who get distracted, drunk, tired, impatient, angry, and just plain overwhelmed.
Makes a lot of sense. Instead of insurance premiums being paid directly by consumers it'd be paid by car manufacturers who pass it on by slight price increases. It could work very well.
I'm pretty extreme, I'd even accept a look-the-other-way policy for the first 10 years where vehicle manufacturers are given a few extra shields from liability. The potential for this tech to save thousands of young lives is very real and we should be bringing it to fruition with indecent haste.
Car accidents cost over a million lives worldwide every year - 45,000 in the USA - and injure about 40 times that number. They are the leading cause of death for people aged 18-34. In the USA, the total cost of accidents is about $230 billion (1.7% of American GDP according to my calculations).
Robot cars will reduce car accidents, probably by a lot (human errors make up 93% of the causes of car accidents; 80% of accidents are due to human inattention). They'll make parking vastly more efficient (the car can drop you off and go and park in some optimal place). 50 billion person-hours are spent driving every year. In productive hours, that's worth a trillion dollars. Even if the hours aren't productive (watching TV or whatever), it's still worth a lot. Congestion would be reduced because robot cars don't need the follow space between cars that human reflexes require. It's like quadrupling the capacity of existing roads for free. Since the crash rate might be so low, cars might be ultralight rather than made of armor, so they'll use less energy and less metal.
This is the same problem current safety tech already has. We have cars that can stop themselves in most situations already - things like that + lane detection + distracted/sleepy driver detection would do wonders if everyone had them even without full autonomy.
But they also all add cost to new vehicles, and the average age of vehicles on the road reflects that. It's been climbing, looks like 11.5 years now.[0]
Looks like there's about 260M cars in the US, and on average only ~18M new ones sold each year[1]. Let's go conservative, and say we just need to replace 200M cars with current-or-near-future tech ones to save 30K lives a year. Let's say a 30K car has the level of tech we want. 200M * 30K = 6 trillion.
But car crashes are like 2% or fewer[2] of all deaths in the US. So that's a massive cost that might be much better spent elsewhere. It's not that we can't make cars safer - we can already do that! - it's that there are some much more appealing ways to spend the money if approaching it from a non-commercial perspective.
We will. For consumers like you there will be a $100k option to increase safety by adding sensors/compute power or something else on top of a base model. But the ultimate game here is to free people from the need to drive cars for a living or pay someone to do so. This will also open a lot of other optimisations to human life, which exist now due to cost of human labor. If human stays in the car - this mission is failed.
+
It's not like if you are the most careful driver you are safe - if a drunk no-licensed jerk aims for your car - you are dead due to his lack of driving ability, not yours.
Why would fewer deaths be impossible to quantify? Insurers already lower your premiums if your car has modern safety features, like automatic emergency braking.
I'd posit that if we replaced human drivers with Tesla's FSD (never mind Waymo tech) today, we'd save lives on the net. We just accept that getting T-boned by a texting driver is "an accident" and those "just happen", whereas getting hit by a self-driving car is "omgwtfbbq those self-driving cars can't be trusted", so standards for self driving cars are unnecessarily higher. (Just imagine the median driver; by definition 50% of drivers are worse.)
If the problem is millions of traffic accidents from individual human drivers then it will be far cheaper and faster to bet big on public transportation, high speed/light/etc. rail, instead of a fully autonomous self-driving car future that has always claimed to be just 5 years away (for the last 15 years).
To successfully develop self-driving cars, without spending NASA moon mission like time and money, we need to put a value on a human life. Maybe even different values for passenger, driver, pedestrian, other motorists, policemen, etc. Then the risk, time, and money to spend will be easily defined. If some task will take a software engineer (paid $150k/year) half year to develop but only increase safety enough to save 1 pedestrian per decade, and a pedestrian is valued at $50k, then maybe the dev can work on something else. Numbers are made up just to illustrate the point. But this is how Tesla, or any other self-driving car-maker can quickly bring self-driving tech to the public in a cost-effective manner, and within reasonable time-frames.
Average human is actually quite terrible in driving cars, as evidenced by the global death toll numbering in dosens of thousands every year. Even halving this number would be a win, losing a lot of people to such edge cases. However, the nature of the problem would not allow for it, because every death from a self-driving car will get excessive coverage and public outrage, and people will prefer losing far more people to human errors than fewer to computer errors.
Another thing that gets missed is this: With self-driving cars every single fatal accident will be investigated, and the result of that investigation will be used to improve the software on all cars in the network.
In this way the rate at which cars kill people will decline, similar to the systematic decline in airplane deaths over the years.
So maybe it's worth initially accepting a self-driving fatality rate that is slightly above the human-driving rate?
Approximately 35,000 people died in car accidents in 2015. That's almost 100 people each day. Imagine you had self-driving car that was exactly as good as humans. This car will kill 100 people each day despite of being human-level competent. It has slim chance of surviving public outrage and lawsuits. This points to the fact that self-driving cars needs to be few thousand times better than humans. As in all machine learning problems, as you get along the curve, progress becomes much more expensive for next 1% improvement than last 2% improvement.
One way to circumvent this is establish special lanes for self-driving cars where things are much more controlled, well defined and cars in that lane can communicate with each other to avoid crashes. Long segments of highways might be great candidates. This can heat up the virtuous cycle where people buy self-driving cars to be in that lane which pressures authorities to make more lanes available for them and eventually most lanes are for self-driving cars.
What if we could make every car $5-10k more expensive instead?
</satire>
I once thought that using AI to make all cars self driving might be the key to making pedestrians and cyclists safe. Self driving cars are playing a game of Russian Roulette. The systems will get it wrong occasionally, with LIDAR or without. Not if, but when. Whether or not someone dies depends on the situation.
reply