Large birds don't fly by flapping their wings when they have enough lift. Look at one in a thermal updraft sometime, it's quite a majestic sight.

If birds had engines on their wings, they would probably fly more like airplanes.

But cameras are a poor substitute for human vision, because they can't move or pivot or refocus much.

It does seem reasonable, though, that if humans can solve safe driving with a brain mass (3 lbs) that computers wouldn't need that much more help. Put another way: every machine learning "expert" who has a kid is amazed to learn how simple unsupervised learning is and how few training examples are required.