Currently AI’s laws have a lot of ways to interpret them. But I do have concerns over a particular way I have seen them interpreted recently:
AI sees crew member gunning down other crew
AI hears sec call to arm
AI bolts armoury and turns off power
Sec arming would cause crew harm
Im going to complain about 3 things. The idea that preping for this is powergaming, what the interpritations will lead to without just arbitrarily changing it mid-round, and lastly how its un-fun.
Preparing for the situation being “powergaming”:
If a secoff knows that this is even possible, whether they have it happen to themselves, or hear about it happening on other stations, why would they ever risk leaving AI in control of security doors? You would cut the AI wire immediately, or otherwise disable the AI the moment any situation comes up that even COULD lead to arming.
You cant compare this to power gaming a malf AI or the like because its the base laws. This is something AI are known to do, prevent Sec from arming up and protecting crew.
The Law Logic:
If an AI thinks forward enough to know sec will harm the shooter with the weapons they are getting, then AI should also know that the shooter will harm people with the weapons they already have even more if you go by a law of numbers view on harm, you would assist the Secoffs.
Ignoring law of numbers, but taking this to an extream, someone is trying to tesloose. Sec are trying to get in to stop them. Sec are pre-armed and at the door to engie, coming for this tesloose guy. AI with this logic on laws should bolt all doors to the tes-looser because sec will shoot them. Thus AI is the ultimate Tesloosing machine. AI will stop sec at every turn from getting a perp because “sec will cause harm”. Why would NT ever make AI that works like this and keep it on the station as apposed to just having another secoff on cams or something?
Or what about AI locking down armory every nukies round because they bought some IDs and sec might harm them?
Lastly fun logic:
Does anyone want an AI that constantly helps antagonists get away with crimes and hampers sec? Do we want every game to start with changing the AI’s laws to robocop because otherwise AI is strictly a tool for antagonists? Do we want every nukies round to auto lose for crew because they all have clown IDs? Does wrestling with breaking down AI’s walls every time before dealing with a wizard with a stolen ID sound fun to anyone?
I honestly think this type of Law Interpretation should be HEAVILY discouraged because it will not lead to an enjoyable state for the game.
Bonus:
Why the fuck is antimov so expensive if with this interpretation it just means AI will go from stopping sec while you attack everyone to attacking everyone itself?