Comment by ux266478

I would pose that's a matter of holding it wrong. Like, integrating generative AI as a prompt to an API endpoint exposed to the user is one of the worst, most naive and bland ways to use it as a game subsystem that I can think of.

Don't get me wrong, there are plenty of problems with trying to use generative AI as a subsystem for a game, problems which exist at a conceptual level. But all the games I've seen that integrate AI as a subsystem don't even come remotely close to hitting that ceiling. I think it's ultimately up to the fact that an interesting usage of generative AI requires a level of technical competency and theory that's well beyond that of the layman. The people making games that utilize AI generally seem to lack the chops to make a good game or do interesting things with AI. The people with the chops and manhours to make a good game with AI:

- Don't want the backlash of having a generative model in their game, because no matter the context that's a lot of very bad press for something that's very difficult and time consuming to implement.

- See the immediate problems and difficulty that arise from a decently integrated concept and execution, and even without the bad press, understand that the juice isn't worth the squeeze. That it's always going to be an immensely stupid gimmick in the end, one that's trying way too hard.

That won't stop us from having 10,000 pieces of shovelware getting churned out that have the worst, most uninspired integrations imaginable mind you. We clearly already have those. For those that could do something better with the tech, there's too many good reasons to not touch it.