That would be a very surprising result to me. I don't think we're maxed out on automation until the ships are loaded at the port, travel to their destination, then unloaded at the destination with no human labor.
Unless you're telling me that robots load, unload, pilot and maintain the ships at sea, we aren't anywhere near maximum automation.
Unless you are willing to risk loosing the ship due to a 1-in-1000 event which your automation did not take into account, then this is your limit. And ships are expensive, even compared to officer salaries. And those generate most of the crewing costs anyways. So if you have to pay 6 officers, adding 2-3 for the engine to require less dock-time and 12 ratings really doesn't push up your costs that much (especially not compared to the capital cost). But you get a lot of redundancy.
So yeah, I would say most of the economically viable automation has been done
But this has been a surprisingly interesting rabbit hole, thanks for that.
There are different numbers for ships around starting with 50k [0] on the low end and the high end is 100k [1] being operated by between 1.5m [2] and 1.9m [3] personell. At the outer edge this would give 40 people per ship. Whereas reports put the number more at 20-25 [4]. The last one also mentions that there are requirements, by law, which typically require something like 6 people (often with nationality requirements) to be on board.
All sources e.g. [0] https://www.statista.com/statistics/264024/number-of-merchan... [1] https://hbs.unctad.org/merchant-fleet/ [2] https://www.ics-shipping.org/shipping-fact/shipping-and-worl... [3] https://www.ics-shipping.org/press-release/russian-and-ukrai... [4] https://www2.deloitte.com/content/dam/Deloitte/global/Docume...