I like to claim that our USVs are artificially intelligent (AI). In the computer field there is some debate over the definition, but I’ll explain our criteria, which are along the lines of the "Turing Test". Essentially, the "Test" is … if a human can’t tell the difference between a human and a working computer … that’s as close as one can get to being a proof of AI.
If you have the opportunity to see our vessels running autonomously you won’t be able to tell whether they are being operated by a human remotely, or navigating on their own. Further, they learn. When they remotely sense an obstacle, they learn to avoid it in the future.
We are also happy to claim that our software is scalable, and completely capable of taking a container ship or an LNG tanker across the ocean safely. However, we are definitely not promoting the use of our systems for unmanned large vessels. We simply call our systems "aids to navigation". To us, while unmanned large ships are feasible … we understand their complex systems and safety requires potential human intervention. We think Lloyd’s would agree.
Certainly, disasters like the Costa Concordia would be made extremely unlikely, unless a captain countermanded the system … which would leave an indisputable forensic trail at a bad captain’s peril.