make AI roguing slightly harder/unroguing easier

[Kotomata]

i think everything is fine as is. i don't really see a reason to nerf going for the AI as it is something you have to go out of your way to do, and if anyone is aware of you trying to break into the chamber, pulling it off becomes incredibly hard -- having the AI be notified automatically when upload has someone sounds annoying. and we have viewports already anyways
the AI has bolted doors, a plasmafire and lethals in their chamber? take down the rwalls with an RCD, tank the fire with a heavy firesuit and use CE's aurora belt to reduce laser damage while you make walls around you.
it's not something impossible, but you may have been approaching it wrong. antagonists / silicons can't interfere either if you bring sec along

obviously you can fix rogue AIs even when they're fortified, that isn't the point
my opinion is that either subverting the AI should require more effort for the bonus it gives you (at no cost of antag power points)
OR it should be easier to de-subvert the lawrack, to lessen the difficulty gap between subversion and de-subversion, and to make the alternative, less fun option of killing all silicons forever* less appealing

also i should have mentioned but this is very much a classic-oriented balance post, i have no frame of reference for how AI subversion works out on RP rounds but I imagine it's much less of an issue thanks to the roleplay rules, higher pop, and lower high-impact antag count

*yes, you can make a new lawrack and use the linker, but that's not super common knowledge + still more effort than killing them + I don't want to write a massive paragraph detailing all of the ways it can fail/be impeded