"However Google has been much more open with their data. Google's automated car was just past 1.3million miles before its first automated accident: a minor fender bender."
However, Google's algorithms are not Tesla's algorithms are not Volvo's algorithms are not Daimler's algorithms are not Toyota's algorithms and so on and so forth. My point is, one developer's miles driven and safety (or accident) reports have nothing to do with another's.
Also, number of accidents and/or resulting deaths don't mean much when the size of the actually deployed fleet of cars is so low. Yes, yes, I know, Tesla is churning out cars like mad, still, their numbers are close to insignificant. When they will have tens or hundreds of millions of cars out there, then their safety/accident numbers will mean something. But, they'll never ever be zero, there's just no chance, unless all cars just stand still in their garages.
My view on this issue is that until I don't see Volvo or Daimler selling cars with full autonomous driving capabilities, and until Google's car division doesn't say the algorithms are nearing applicability, I won't trust any company selling autonomy features. Why? Because Daimler has been working on this much longer and Google has been developing much longer, and until they don't say it's OK, it won't be OK, it's that simple.