It's not the cameras that worry me, it's the whole issue that AI is not remotely intelligent. Image recognition is step 1, and NNs are pretty good at it, but can still be fooled. However taking those images, properly interpreting them with no error, keeping track of temporary occluded objects, unusual circumstances, and then making decisions is not something I trust NNs to do, as there just is not enough training data, despite the valiant efforts to gather it.

One difference between true intelligence and what we have in NNs today is, if you showed an image or two of an object to a person they could identify that object in other orientations and lighting conditions pretty well. To do the same with NNs requires thousands of examples and counter examples to train. This is why with driving how could you have enough examples of strange issues to ensure the car responds safely?

The only advantage I see today for cars driving is they don't get bored or distracted or drunk, but they certainly are not intelligent.