The point is that some parts of games truly don't impact the experience.

Name a game whose settings screen added to the gameplay in any meaningful way. Hell, name a game from memory that didn't have the same settings screen with minor styling variations as every other game you played. What exactly am I missing out on by having a resolution dropdown that was hand coded? Does a bespoke "ultra" graphics slider bring the characters to life in a way that AI just can't? Besides, using AI isn't mutually exclusive with all of the elements working together. My point is that AI, wielded properly, makes that easier to achieve. If you don't have to spend a month hand coding sliders for a settings screen, you can get on with making the actual game. Do you also rail against game engines which provide all of this stuff in a generic way?

I do think that using AI to help with internationalization is fine, as long as you QA it just like any code or content that is created by man or machine. It's a tool, and you have to know how and when to use any tool. You can get plenty of good translation out of an AI if you are using it properly.

Also worth noting that Silksong was verified to have a human translation team, and they never clarified if AI was used. It's entirely possible that today's SOTA AI would have done a much better job. The problem with your argument about translation is you and I have no idea how many games have been machine translated well, and never saw an issue. Because of the way gamers are (frequently unhinged about minor things. Prone to groupthink. Extremely vocal. Absolute in opinions like "There is no acceptable use for AI in gaming"), you just can't stick your neck out and say, "We used AI for our localizations, then had native speakers QA it, and not one had issues." You risk becoming a lightning rod for puritans because you wanted to give Romanians a chance to play a game in their own language.

Even if you disagree on using AI for these things, I can keep coming up with ways that AI can make a game better. That's why I told you not to argue against specific examples. Maybe a game designer told an AI to review the code and find performance improvement opportunities. Would you be upset if the game was half the size, and could run higher frame-rates with the exact same content? Is that still an unacceptable us of AI, or would you rather that the man hours that could otherwise go to things that humans are better at get spent on optimizations. I personally would rather have a game with better performance from an AI and more human created side quests, than a game with better performance and fewer side quests because they ran out of budget hiring people to dig into compiler settings.

So yeah, I think it is completely reasonable to use AI in gaming or any domain of programming. I've been around long enough to remember human teams turning out absolute steaming piles of garbage. The issue isn't AI, the issue is people turning out steaming piles of garbage. AI just makes that easier, but by no means does it make it inherent to use.