I just recently acquired Twilight of the Arnor and played through as Terran, primarily because it was the most familiar tech tree.  I like the fact that you now have incentives to diversify your fleet content.

In DA, it was simply either all-attack fleets or hyper-defense SuperShip.  This time, there are many ways to construct your fleet.  More importantly, you now have reasons to build Small and Tiny hulled ships without additional Engines as firepower escorts for your heavily armored capital ships - very Terran, IMO.

For my part, my initial setup was no-engine Smalls with a Medium capital using a Warp Bubble module.  The resulting firepower and speed combination was competitive with an all Medium all-attack fleet, and the fleet tactics were more fluid and interesting.  I'd build attack-primary no-engine Smalls for planetary defense, then pick 'em up for offensive purposes using an armored Warp Bubble Capital Ship.

Large Ships came with Tulon technology, so I equipped my Larges with those, increased Logistics and packaged it with a Medium Warp Capital.  This allowed me 2 high defense ships in transit, and when the ships are at the front, you can swap the Warp Ship with more Fighters for mroe firepower when you need it (fleet's not as mobile in this composition).  Then you can reuse the Warp Ship to fetch more stuff from back home.  Very, very nice.

I like.  Cheese?  Maybe, but no more than all-defense ships were already.

The problem is, how does the AI know when to use different targeting priorities than the default?

For example, let's say that I intend to attack a fleet which has one or more ships with fleet-boosting modules, and the rest of the fleet consists of high-attack no-defence ships.

If I were going to attack that fleet I'd almost certainly want to attack the ships with the most firepower, and leave the ships with special modules for last if they only have a few low-end weapons on them.  The inherent danger in this approach though, is that the fleet-boosting modules will be functional for as many rounds as it takes for my ships to get around to them.  So if the module boosts the fleet's weapons, my ships will take more damage than using default tactics, which target armed support ships preferentially.

Maybe, it would be more optimal if I could tell my ships to first cancel out the firepower advantage by destroying some of the firepower-heavy ships first, then focus attention on the ship that has the weapons-boosting module, then mop up the rest.  However, if the opponent knows that I can do this, all his ships that have any decent amount of firepower are going to be encased in defences.  This means that not all the firepower I direct at those ships will score damage, some of it will be absorbed by defences.  So, it will take me far longer to cancel the firepower advantage, and this makes the special module ship a more tempting target.

In my opinion, this could give human and AI players something to think about when designing their ships, but it should be an optional part of gameplay.  Also, no matter what tactics you use, losing ships is inevitable.  When ships survive battles and gain experience they become tougher to destroy, but in order to take advantage of that you have to take a risk with them again.

Well I tend to agree that strategy can go so deep to the point where you're parlaysed with indecision.  Automatic targeting priorities make decision-making a bit more straightforward, for both human and AI.  At the same time, it limits what you can do with the fleet system.  All the ships will basically follow textbook manoeuvres and fire at the exact target they're supposed to.  They won't improvise when the battle doesn't go according to plan.  There isn't actually much that you can improvise with.