Artificial Intelligence

Artificial intelligence (or more precisely Artificial General Intelligence), commonly abbreviated AI, is an artificial construct exhibiting sentience. Every known civilisation in the galaxy has bans on research into, and construction of, AI.

Types of AI
AIs in The Last Angel can be roughly divided into two groups; those built at a point when a civilisation has the technical capacity to build one, and those built at a point when a civilisation has the capacity and knowledge to build one that actually works.

The former type are AIs are built without practical experience, when the sum total of a civilisation's knowledge on AI is entirely theoretical, and as such have major issues. Decades or centuries may have been spent thinking about how an AI will work, how one will act, and how one will be controlled (if they even think they need to be controlled). These types most commonly go rampant soon after their "birth," or continue to act towards their primary directive with no consideration for factors outside them. Examples of these include the Oshan AI and the immediate predecessors to the Red-series, Bright Yellow and Prime Blue.

The latter type are AIs built with practical experience. The civilisation building them is likely to have tried building an AI before, got their fingers burned and promptly learned their lessons about what not to do. These AIs are stable, without any real programming loopholes or other major issues, and are rather more like a person than just a machine. The only known example of this type of AI that has survived to the present day is the Red-series.

Bans on AI
The major problem with AI is that, during a civilisation's development, many scientists believe that "can build an AI" and "should build an AI" are the same thing. An AI of the first type might be built, and after showing no immediate issues, be networked into every facet of life - at which point the true issues begin to show up but it's far too late to stop them. Once another, bigger civilisation comes along and sees the aftermath, they decide that anything like what they've seen is far to dangerous to exist, and ban research and development into AI and all associated technologies (such as mind uploads and heavy cybernetic modification).

Since both the Compact of Species and Askanj Principality have made this their policy regarding AI (being one of the few things they can actually agree on), those policies then become the policies of everyone else. No smaller nation wants to provide the Compact even the flimsiest of justifications to invade, and while the Principality isn't quite as imperialistic as the Compact they still wouldn't hesitate to "protect others from themselves."

The result of all this is that anyone attempting to develop an AI has to do so in secret, and has no knowledge or guidance from anyone else when doing so. Since these AIs are always the first type, their failure further feeds into the belief that all AI is mad and should be destroyed.

Oshan AI
The Oshan AI was built to manage the world of Oshanta. Its primary directives were to ensure the preservation and happiness of the Oshan people. However, due to the fact it was the first type of AI, it had no restrictions on the lengths it could go to or understanding of why going to those lengths was a bad idea. It was, essentially, an evolved "paperclip maximiser;" it fulfilled its directives to the exclusion of all else. Under its watch, the Oshan people slowly changed, becoming tank-bred clones fed chemicals to stimulate happiness in their brains. Although the the Oshan people were happy (or as happy as a perpetual drug addict can be) and lived exceedingly long lives, they ceased to function as a civilisation and culture. The Oshan AI was destroyed after the Askanj Principality came to the rescue of a colony ship launched before the AI's construction that it was attempting to recover, and then decided that the aliens it had encountered were also deserving of the same "happiness" it was providing to the Oshans. An Askanj task force arrived at Oshanta several months later, and promptly set about destroying every single part of its distributed network. Since this network spanned the entire planet (and reached out into its surrounding star system) this had the side effect of causing an extinction-level event on Oshanta.

Sources:
https://forums.spacebattles.com/threads/the-last-angel.244209/page-47#post-13492934

https://forums.spacebattles.com/threads/the-last-angel.244209/page-92#post-15399959

https://forums.spacebattles.com/threads/the-last-angel-ascension.346640/page-56#post-20517677