The technical investment to produce a fully automatic air transport system - not just the aircraft but all the operational stuff that surrounds it - would be considerable. If done from scratch but adapting as much as possible from existing airframes electronics, procedures, etc, the level of develpment effort might be somewhere between 50,000 and 100,000 man years ..maybe 8 to 15 billion dollars in current money. That's a lot to swallow.... especially for a very slow payback..if any.
The maintenance and support costs associated with qualifying and operating robotic aircraft are not so clearly better. Lots of technicians needed in the loop. Not cheap. And harder to manage. Much less disciplined than aircrew meeting contemporary standards.
Chances are, pilots will be more economical than the whole robotics bit for a long time to come. Liability is part of it... customer perception.. economics. None of these are screaming for full automation.
Likelier is some gradual evolution from more 'handling' to less, and gradually increasing the population of crews and planes ok for cat3 type operations. Automation that reduces tech stops and increases successful arrivals in dicey weather will pay for itself much sooner than the full automation deal.
Good pilots are a bargain. Always have been.