Hard to see the pedestrian in those lighting conditions. As you say, difficult to tell whether a normal driver would have managed to avoid her. You would actually have thought that the computer might have stood more chance of detecting her with microwave radar or some form of night vision.
Certainly the pedestrian shouldn't have crossed there.
That's part of the point really. These systems are touted as being better than the average driver. Not affected by weather conditions or poor lighting, etc.. The system recognized that there was something in its path and failed to respond accordingly. The question now becomes at what distance did it recognize that she was there? Did it recognize that she was moving and her direction of travel was likely to intersect with that of the car? If the system(s) can only "see" as far as the human eye in similar conditions, what is the point of them? An operator of a vehicle is responsible for operating the vehicle in a safe manner and not driving faster than is safe for the road conditions. The posted limit is the safe limit for ideal conditions. If the automated system doesn't allow for rain, snow, dark, etc.. and drives the posted limit regardless of road conditions, it isn't safer than the average driver. As for the pedestrian shouldn't have crossed there, well people do unpredictable and stupid things. If it was a regular car, the road conditions, lighting, and a myriad of other factors would be taken into consideration as to the level of culpability of the driver, I'm sure. But the driver had relinquished control to a system that is supposed to be at least as good as him. It didn't even react and slow down.
I'm not a complete Luddite, but I'm not a crash test dummy either. As long as people remain unpredictable these things should remain off of the roads.
Steven