You can define it however you want. By any common sense definition of safety Tesla has not proven that their autopilot is 'safer' than a human driver.
>10x or 100x the number of deaths would be small price to pay to get everyone in autonomous vehicles.
This supposes a couple things. Mainly that autonomous vehicles will become safer than human drivers, that you know roughly how many humans will have to die to achieve that and that those humans have to die to achieve it. Those are all unknown at this point and even if you disagree about the first one (which I expect you might) you still have to grant me two and three.
Can Autopilot not run passively and prevent human errors? Why are the two options presented seem to be only "only a human behind the wheel, not assisted by even AEB" and "Full autopilot with no human in the car at all".
In that study there are two buckets, one which is total Tesla miles in TACC-enabled cars and then after the update total Tesla miles in cars with TACC + Autosteer and they calculated on airbag deployments. Human driven miles are going to dominate both of those buckets and there's a reason the NHTSA report makes zero claims about Tesla's safety relative to human drivers. It's totally outside the scope of the study. Then add in that some researchers who are skeptical of the methodology and have been asking for the raw data from NHTSA/Tesla have yet to receive it.
https://news.ycombinator.com/item?id=16932350
Autosteer, however, is relatively unique to Tesla. That’s what makes singling out Autosteer as the source of a 40 percent drop so curious. Forward collision warning and automatic emergency braking were introduced just months before the introduction of Autosteer in October 2015. A previous IIHS study shows that both the collision warning and auto emergency braking can deliver a similar reduction in crashes.
https://news.ycombinator.com/item?id=16932406