There's less air traffic now. But if you approve the first autonomous air taxi, you'll soon approve the second and third and before you know it there are thousands of those things whirring through every major city and then you have just as much traffic and one more dimension to worry about.

Agreed, above trees and buildings there is a lot less air traffic to worry about. But you get into the inherent dangers of air travel. Helicopters are especially dangerous, unlike planes if they lose power they cannot glide at all. In addition they take off vertically, assuming there will be set takeoff landing areas, checking for rapidly ascending and descending aircraft will be very important. Birds are always a concern when it comes to propellers too. And if used in a city up and down drafts created by large buildings like skyscrapers will provide a large controls problem, let's hope those controllers can reliably handle impulse forces.

Specially if things are built from the ground up (pun intended). A new system relying in communication between software and sensors should be relatively easier to deal than the fuzziness of reading signs and reacting to random elements around you.

I hope they have excellent navigation system which at least won't crash the aircraft if the gps/glonass/etc signals suddenly got disrupted (bad weather, interference, military activity, etc). Having a big taxi drone suddenly trying to emergency autoland on your roof due to gps failure would be horrible.

I can see how the autonomous control part might be simpler due to there being fewer objects to avoid colliding with, but there’s the no-small-matters of the additional dimension to navigate combined with managing complex avionics vs the simpler control mechanisms of a car. Dealing with takeoff, landings, crosswinds, and many other things are much more complicated than driving a car.