In defense of the developers… crafting artificially intelligent behavior for games like this is pretty tough. The complexity of these ships’ virtual systems is a lot to contend with; imagine writing detailed instructions for someone else to exactly mimic your gameplay style and habits! It would be difficult enough to explain just how to operate the ship’s systems, let alone how to choose contracts, augment one’s ship, etc. It’s much easier in things like a FPS where non player objects have a fairly limited set of possible behaviors and interactions.
It’s also not enough to write instructions to make the computer player smart; they need to be controllably smart. If one were to write the behavior of the AI in a FPS to be as efficient as theoretically possible given the game’s engine, the computational resources available, etc, it would likely result in a fairly simple program. However, human players wouldn’t stand a chance against computer players that could hit them from across a map with 100% accuracy at the very first moment there was a clear line of sight between them. Again, the relative simplicity of the FPS works to its favor here as it’s easy to tailor the difficulty with simple tweaks to things like enemy health, accuracy, enagagement range, etc. While there are some similar attributes in OiS that could be similarly tweaked, these don’t begin to cover compex behaviors, like system management, or those that require abstract tactical thinking, such as utilizing EMCON mode.
Another problem is one of computational resources. Let’s assume one developed a very sophisticated model for non-player behavior. Perhaps it could be easily adjusted to have different play styles, skill levels, agression levels. One could even leverage all of the great new machine learning and AI tools and libraries available now to develop something so sophisticated as to seem almost human. However every time the computer player makes a decision, each bit of complexity translates to more clock cycles. The bigger the progam is, the more memory it takes. The more non player ships there are, the longer it takes for a single cycle of the game’s physics engine to run. Fairly soon, one can outpace what most computers can handle, and then the game is effectively unplayable. So, developers also have the challenge of not only designing AI that can be interesting, varied, clever but also making to be feasible for the lowest end macines they support to run.
The nice thing about all of this, though, is that AI is something that can be improved over time after other parts of an application are stabilized. It’s probably more important that they focus on ironing out bugs and making all of the other features of the game as polished as possible before putting more energy into the AI scripting.
Just my two cents…
tl:dr; Writing game AI is really hard. This is a really hard game to write AI for. Stabilizing the game first and ironing out other bugs will make the AI easier to write. Be patient and ye shall be rewarded.
DISCLAIMER: I am not a game developer, but I am a developer. I have worked in many different disciplines of SE/CS including some dealing with the same concepts that games are built with.