well, maybe if instead of actual artificial intelligence (aka something which knows and MAY learn) he refers to simply a inside automatic response, then it's an ai.
to say.
set the "ai" to open the bridges in intervals when a siege is there.
siege is when tot units walk upon tot pressure plates and destroy them.
when that happens also open the magma channels.
it's an automated response, which could be fine and doable.
now, to that you add that, when this happens, the displays shows off "siege incoming" "deploying countermeasures" "countermeasures deployed".
it's not an ai, but it feels like that.
it's the difference between making a program which fetches data from a txt, and a program which fetches data from himself, but making the one who runs the program read on screen "program fetching data" it's the same stuff, but in two different ways.
U_U
and it would still be a hell of a work to make.
now, maybe it's an "ai" in the sense of calculations OR display showing predeterminate letters.
like "press the lever to ask "everything fine?"" answer will always be:"yes it's fine, more magma".
faking an ai is still a hussle, and if done well it's quite difficult to realize it.
since after all, a program who can learn by himself does not, quite actually, exist.
(there may be similar programs, but that's different, it's simply storing data in relation to words, which is what clever bot is)
to say, the program should learn by himself, without being asked to learn.
he should thus feel egoism, and have an ego.
he should be able to reason for abstract.
and then it would be an AHI. artificial human intelligence.
humans think abstract, have long term memory (pc's have it on a base), and are able to prove curiosity and discern themselves from someone else.
U_U