BioShock creator Ken Levine has expressed skepticism regarding generative AI, admitting its usefulness for technical tasks but doubting its ability to "tell [him] a really compelling story".
In a new interview with GamesIndustry.biz, Levine is asked for his opinions on generative AI and its "capability to react to players". Levine says he doesn't want to "underestimate" the technology, but that its biggest limitation right now is "persistence".
He gives the example of ChatGPT's video generator Sora, which he says is capable of generating a "beautiful" street scene featuring a woman walking down the street. However, if that woman were to "turn around and walk backwards", Sora wouldn't "remember where she has been".
It's this lack of persistence that Levine says means AI can't tell him a "really compelling story that has a three-act structure"; in fact, AI can't "even tell [him] multiple scenes", he says, going on to say that he's "not overly impressed" by the technology as it currently stands.
Despite that, Levine acknowledges that generative AI has its uses. He says that it's useful for "training your bug database to query how many bugs you have in certain situations", for instance, or "clearing [the Judas dev team's] analytics database".
However, he hasn't used it for concept art, largely "because there's some legal issues" around how generative AI sources its images. In case you're not aware, generative AI can often use art without crediting the original artists when generating content.
The full GI.biz interview is well worth reading, especially if you're interested in keeping up with Levine's work on Judas. He discusses everything from the game's approach to player choice all the way through to his work on BioShock.
Levine isn't the only person in the industry to express skepticism when it comes to gen AI. Back in July, Nintendo declared that it had no plans to use the tech in its games, acknowledging similar IP rights issues to Levine.
However, that hasn't stopped companies like Netflix, Ubisoft, and Square Enix from going all in on generative AI. It looks like this is a debate that isn't going away anytime soon.
