The engineer does all the work and the others are just there to support a tablet and a cup of coffee.

What about piracy? How are they going to deter or prevent physical theft or hijack?

At what speed does a submarine use less energy to move than the equivalent carrying capacity surface vessel? Should the robotic vessels all travel submerged?

I agree with Domain Rider; My first thought was also, 'What about piracy? How are they going to deter or prevent physical theft or hijack?' Maybe amazed W1's query regarding submarines is the answer, they wouldn't need to be very deep, but trying to hijack a submerged vessel would be very difficult.

Shipping companies are looking to cut costs, what do you think this will go for, versus a trained crew?
Besides those drones won't deployed during heavy winds, fog, storms or millions of other reasons, least of the pirates already on board...
By the way that strange noisy you think you hear is a torch cutting an opening inside...

With no hostages available and a sealed hull, it would be pretty hard to hijack.

This ship scares the hell out of me. I'm a crushing sailor. A ship like this can come up on you at night, quiet and invisible, and never sees you. With no human witnesses, you never existed and now you're shark meat.
I don't see the slightest consideration to sailboats on this ad. There are no traffic lights at sea, RR! There is no way to avoid shipping lanes on a sailboat!

Other hazards not considered, silting in harbours, piloting through narrow channels etc - this is not realistic at all. We have not progressed technology far enough to enable full automation - I know this as I work in marine automation. Tidal currents etc are also an issue, as well as anyone at sea understands - no matter how large the vessel, you can have the correct heading and still end up miles off course. You can never replace human judgement as so many accidents at sea are already attributed to navigational auto-pilot systems and the heading/course issue. Rolls Royce have made a lot of far fetched assumptions here - but she looks nice :-)

Too many problems for 2020, including issues like piracy, equipment failure, etc. Maybe by 2050 though...

I am amazed about all the concerns over piracy. Very few ships gets attached by pirates. It's just been in the news lately. With no humans on board, it will involve less exposure, insurance and risk in case of attack by pirates. A good thing. Also, who says an autonomous system will be less able to avoid pirates than a human crew? The ship can be remotely monitored. Launch some drones and take those pirates out (many ways to do that without killing them).
This tech is in relative infancy, but it is the future. In 20 years it will be well on the way to going mainstream. Ships have been getting bigger and crews smaller. This is the end goal of the trend. People building wooden ships were making the same short-sighted assumptions and comments about the first concepts or steel ships. Or sail boats vs steam boats, etc. Horse carriages vs automobiles. Just like people currently are nay-saying the impending change to renewable energy and autonomous electric cars. Electric cars has had many false starts. 'Proof' that it won't work. Its all about battery cost and energy density expressed in $ per kw/hr. We will be there in 5 to 10 years.