Fortunately for us t's more complicated than this. Unlike most other areas of AI and automation, failure here will be spectacular...and deadly. As another poster alluded, you also need major advances in datalink waveforms or bandwidth. Even the USAF can't provide all required bandwidth organically - and they own and operate leading edge space based comm. Then - even if you solve the 'throughput' problem you still have to make it secure. Not just secure like your facebook password, but REAL security. Another poster did a fine job explaining the basics of what that means. When you even get to this point, you need redundancy for normal material and infrastructure failures. Now tell me WHO is going to pay for this? The cost to build and operate the required infrastructure will be staggering. I know that we make good money, but I don't think you'll find any cost savings for a very long time.

The lessons learned from US military UAS/RPA are instructive. The control systems and communication/control links are arguably more important the the technology in the aircraft. In my experience the systems lacked adequate operational redundancy and flexibility for commercial operations.

Based on my 121 and military experience I think that introducing these systems into the airspace will only slow things down and reduce capacity. Ask yourself how many times you've saved a go around by having the preceding aircraft in sight. How many more aircraft can arrive under 'visual' conditions - and how much do airports rely on that capacity?

I'll believe it's coming when I see more efficiency & no mistakes from ATC, when I see no ACARS, WiFi, & CCS failures, and when I see no Jepp Database or aircraft systems malfunctions.

In other words - I don't expect to see this in my career, or even my lifetime. FWIW I'm 53 years old.