I think you're underestimating the complexity of the machine. To me, the navigational stuff would be the easier part. The big void would be left on the engineering side and the "state of the art" automation in place now isn't even all that reliable. I have yet to work on a vessel whose "automation" system didn't have numerous glitches, false readings and failures. Some were better than others. Nice to glance at, but not the end all, be all. Never worked with RR but everything I've heard regarding their marine systems is nothing but nightmares.
I think there is a widespread mis-understanding of how technology gets introduced and how it changes things
First its much more iterative then people think , there is almost never a "instant" change , often the media like to present it that way ( the xxxxx revolution ) , but in reality its a slow iterative process that often has been going on for many years out of the public gaze.
So today, we have a progression to increasingly automated and integrated bridges, automated engine
diagnostics and monitoring of engine
spaces ( see super yachts )
I could easily see a future where the piloting is done remotely like drones , with a grease monkey or two to handle minor issues.
Today we see greater and greater pressure to economise on sea freight costs, de-manning has gone along way, but the pressures will continue to remove more and more shipboard personnel.
Will be see unmanned ships in the next 10 years , no , but we will see a steady progression to more and more automated ships.
The fact is if we can run robotic cars and unmanned mission to mars we can certainly get a ship across an ocean.
Think fully enclosed cargo spaces, redundant engineering systems, diesel electric
multiple engine pod based propulsion
, etc etc. you just design the systems to handle failures andn keep going.
Think of the advantages in concentrating engineering skills in ports
, rather then on the ship itself.
When the day arrives and the ship is finally unmanned, we will not blink an eye. !