The carriers are what help you win that battle, I think the AI should make carriers a high priority target.
Yes and no. I don't think it is so simple. It's not which ship is a higher priority target, as that changes, but rather which is currently of high value to attack. You're right, to win my enemy could have forget my cobalts and attack the carriers, but if it always attacks carriers first it creates other problems. Moreover, they needed to be able to kill the cobalts, not chase indefinitely.
Major- I know, but the dev's have also commented that they wish for us to post examples of such shortcomings for them to better improve. Moreover, my post more than simply to show said shortcomings, but rather illuminate the large problem of the advantages of micro-managing battles, which I guess is partly a function of the AI. So, yes, as the AI improves, this advantage will decrease, but I doubt it will diminish completely.
So far as I can tell, based on watching my own fleets' AI, the ship AI and tactics are those of individual ships performing individual tasks, often with individual goals in mind. This is what leads to my advantage when I manage my battles. I can use the individuals to perform group goals: defensive positions, focus fire, and flanking maneuvers. It is my opinion, these types of goals need to be recognized with respect to how to improve AI. So yes, there are many other threads, but none that are currently actively talking about this.