Teslas crash 40% less than other cars, and 1/3 the number of people are killed in Teslas versus other cars.
Indeed once a common failure mode like this is identified it needs to be investigated and fixed. Something similar happened a few years ago when someone driving a Tesla while watching a movie (not paying attention) died when they crashed into a light-colored tractor trailer directly crossing the road. So an investigation makes sense. But much of the general criticism of self-driving and autopilot here seems misplaced. Teslas and other self-driving vehicle technologies are saving lives. They will continue to save lives compared to human drivers, as long as we let them.
I really wrestle with this line of reasoning. Tesla keeps pointing this out, and it's appealing to me, but at the same time something about it seems off to me. I can't tell if this is erroneous intuition on my part blinding me to a more rational assessment, or if that intuition is onto something important.
Some top-of-my-head thoughts:
1. I think to make a fair comparison of Tesla versus other cars, you'd have to really ask "how much safer are Tesla owners in Teslas compared to other cars randomly assigned to them?" That is, comparing the accident rates of Teslas compared to other cars is misleading because Tesla owners are not a random slice of the population. I almost guarantee that if you e.g., looked at their accident rates prior to owning a Tesla their accident rates would be lower than the general population.
2. In these autopilot situations, bringing up general accident rates seems sort of like a red herring to me. The actual causally relevant issue is "what would happen in this scenario if someone were driving without an autopilot?" So, for example, in the example of the rider who was killed when the autopilot drove them into a semi, the actually relevant question is "what would have happened if that driver, or someone interchangeable with them, was driving without autopilot? Would have they drove themselves into a semi?"
3. Various experts have argued general vehicle accident rates aren't comparable to Teslas because average cars are much, much older. As such, you should be comparing accident rates of cars of the same age, if nothing else. So, aside from the driver effect pointed out earlier, you have the question of "what would the accident rate look like in a Tesla or a car identical to it without autopilot?"
4. At some point with autopilot -- whether it be Tesla or other companies -- you have to start treating it comparably to a single individual. So, for example, what are the odds of Person A27K38, driving the same number of miles as Tesla, having a certain pattern of accidents? If you found a specific person drove into first responders on the side of the road 11 times, wouldn't that be suggestive of a pattern? Or would it? It's not enough to ask "how often do non autopilot drivers drive into first responders on the side of the road", it seems to me important to ask "how often would a single driver drive into first responders on the side of the road, given a certain number of miles driven in that same period?" At some point, autopilot becomes a driver, in the sense it has a unique identity regardless of how many copies of it there are? Maybe that's not right but it seems like that is the case.
Teslas crash 40% less than other cars, and 1/3 the number of people are killed in Teslas versus other cars.
Indeed once a common failure mode like this is identified it needs to be investigated and fixed. Something similar happened a few years ago when someone driving a Tesla while watching a movie (not paying attention) died when they crashed into a light-colored tractor trailer directly crossing the road. So an investigation makes sense. But much of the general criticism of self-driving and autopilot here seems misplaced. Teslas and other self-driving vehicle technologies are saving lives. They will continue to save lives compared to human drivers, as long as we let them.