I believe the first generations of driverless cars have to be overly cautious. If it can't be 100% sure that it's model corresponds to reality, and that nothing

I believe the first generations of driverless cars have to be overly cautious. If it can't be 100% sure that it's model corresponds to reality, and that nothing is going to jump out of the bush, then it has to stop. A safe car is going to stop or slow down at least once per trip, for the forseeable future. I think the car should also talk about what it thinks, to get people to understand the limitations of the technology. "There could be a child between those parking cars, slowing down" "I cannot see the road markings, emergency stop" "Road signs don't match my data, slowing down" "Train mode activated, following virtual track" Then, and only then, you can do the second step. Automated cars have the potential to be much safer than cars with human drivers. You can accept some accidents, in order to prevent many more. But you can only do that if the public and lawmakers have confidence in the technology. Right now it is way too much of a black box with wierd failure modes.