Whilst there are some doubts about whethe r the technology is currently as good as it needs to be be for driverless cars, it seems this particular case in America does not seem to tell us anything new based on what we know so far about it. As others have said the accident would probably have happened anyway even if there had been a driver. (In fact there was a driver-but I can see an argument that as they were in a sort of standby mode maybe they weren't as focused as they might have been if they had they been driving.
These trials should continue in my opinion. I am confident that in time they will be safer, quicker and less polluting than cars with drivers (a fair proportion of whom now are not safe to drive- eg use phones, drive over the limit, drive without passing a test, speed unsafely, drive too close...).
We are not ready yet of course- these are just trials. Important problems to overcome at the moment are the inability to detect pedestrians and small objects in the road, and the interaction between driverless cars and human driver cars. It may be 10-20 years before we are ready but it already works for trains and planes so it may become a reality on the roads..