Here we go again! Took it upon myself to start this topic, as I believe it's a pretty big part of the game's design right now and it's also something that can and should generate a lot of long posts. I don't do a whole lot of coding but am reliably informed that silicon.dm code is a nightmare which has tendrils in far more than it should - I do not intend to discuss that but instead discuss this simple "patch-over" that stops anyone joining as AI. This does not touch borgs, which are a related, but different kettle of fish. I've heard lot of people don't like what the AI in its current role brings to the table. Complaints I feel are in three general areas. I will add examples of where I have seen this happen in rounds (all a while ago) 1. Validhunting AI is a validhunty role by its nature. It has eyes over the whole ship, recieves the job description of whoever is speaking and can follow people by ID card. Its laws also compel it to protect the ship and everyone in it thus "laws to their logical conclusion" find valids. Examples (from a non-antag perspective) : the AI telling me someone's job was "agent", AI telling me someone's using a voice changer (because job did not match name), AI following players to tell me what they're doing (known traitor character e.g. "x is standing outside AI upload/telecoms"). Even if the AI is not telling people that it's doing this, it can watch invisibly until you commit a crime then rumble you to security silently. From an antag perspective, I'm forced to do more and more of my evil in maint where it cannot see (also avoids security watching cameras) due to the potential threat that I will not hear until sec rocks up at my door. 2. AI trying to do your job. This is often more of an RP-problem than a mechanical one. AI is expected to know the vast majority of how the ship works, and can offer advice. Sometimes this is useful, but other times the AI will contradict and argue with you about what you're doing. At worst, it can be obstructive and an enforcer of "rules". Examples: AI disagreeing with CO about what access to grant person. AI disagreeing on alert procedure. AI disagreeing with engine setup as engineer over minor matters. AI demanding flight plans over the head of CSO/XO/CO/Pathfinder. I do not see AIs wrestling the controls from people (until they are marked valids in the AI player's mind, see 1.), but I do see argumentative AIs often. Borgs have similar problems, seems to come with the design. 3. Powergaming. As AI, you have some considerable game-mechanical power. Can remotely operate any computer/door/switch and most machines either equal to, or in some cases beyond what crew can do. At worst, it's an omnipresent, all access friend or enemy to anyone on the ship. This is more subjective but a common complaint I have heard. Examples : AI locking a gunman in a room by bolting all the doors shut. AI draining atmos (!) from a room that bad guys are in to kill them (This is probably against the rules anyway, but is possible). Examples that are possible, but I have not seen are bolt/depowering doors to slow people down, ejecting supermatter However, the AI is a role that is integrated into a lot of the ship's systems by its original design. Removing the AI may remove some valuable functionality from the crew of the ship. I'll post a reply to my own thread later to list out some things I feel the AI is needed to do right now, and potential ideas for quick replacements that would need very few new features. It'd be very much appreciated if people could post about things that the AI can do now that can't be replaced by other crew members. Additionally, problems with poor AI play that seem common, and aren't easily fixed by moderators/admins would be appreciated.