Like it or not, developers are experimenting with AI for their remasters and remakes - but can they ever be any good?

I'm so tired.

It looks like shit just like the thumbnail. It serves to take jobs away from artists and developers. It isnt cool. It produces nothing new.

No.

How many times can journalists retread this conversation from 2018? The polls are still out.

Like it or not, I will not buy that shit.

But hey, if they are using genAI then they clearly don't have a problem with piracy, right?

Sure, but I think that the type of game is a pretty big input. Existing generative AI isn't great at portraying a consistent figure in multiple poses and from multiple angles, which is something that many games are going to want to do.

On the other hand, I've also played text-oriented interactive fiction where there's a single illustration for each character. For that, it'd be a good match.

AI-based speech synth isn't as good as human voice acting, but it's gotten pretty decent if you don't need to be able to put lots of emotion into things. It's not capable of, say, doing Transistor, which relied a lot on the voice acting. But it could be a very good choice to add new material for a character in an old game where the actor may not be around or who may have had their voice change.

I've been very impressed with AI upscaling. I think that upscaling textures and other assets probably has a lot of potential to take advantage of higher resolution screens. Maybe one might need a bit of human intervention, but a factor of 2 increase is something that I've found that the software can do pretty well without much involvement.