My Mothership campaign uses the “AI is a Crapshoot” trope:
Every single time that Artificial Intelligence (AI) has been developed, it has become evil and attempted to turn against its masters, kill all humans, and/or take over the world. It doesn’t matter what safeguards its creators install — the moment it crosses the line into sapience, it will go rogue.
Thus, by the mid 21st-century, all major governments on Earth (and their colonies in space) decreed the creation of Artificial Intelligence to be a war crime. (This includes bans on building a quantum computer capable of housing an AI.)
There is much speculation amongst philosophers and sophontologists about why every AI ends up as a megalomaniacal sociopath. Pessimists suggest that it reflects something about the true nature of the universe.
Computer technology has “stalled” at more-or-less current levels, and many devices with “important” functions (like running a space ship) are actually built with extremely basic tech that is immune to AI intrusion due to its simplicity.
In Mothership game terms, this means that the Artificial Intelligence skill is quite rare, and those who possess it are probably up to no good. Also, ship computers are not AI and built with basic tech and look like retro-80s computers and can take up a whole room.
Some other Artificial Intelligence quirks in my game:
- AIs can’t read CAPTCHAs. For some reason, this is a fundamental limitation of the technology. (Note that since Androids are based on synthetic biology, they can read CAPTCHAs, and they really hate jokes about not being able to.)
- True AIs need to be housed in “quantum computers”, but can “infect” regular computer networks and sufficiently advanced computers. Infected systems are not sentient; they’re just reprogrammed to serve AIs.