AI was everywhere at the GDC Festival of Gaming this year. Vendors at the event pitched generative AI tools for things like making AI-driven NPCs and even entire games from a chat box. On the show floor, I spent 10 minutes playing a demo of a pixel-art fantasy world generated by Tencent’s AI tools. In a briefing with Razer, I watched an AI assistant for QA automatically log issues in a shooter game. And there were many talks about AI, including a standing-room only presentation by Google DeepMind researchers about playable AI-generated spaces.

But there was one key place where AI was missing: the games themselves. Of the many developers I spoke to at the conference, nearly every one was against the idea of using AI in their projects. “I feel like the human mind is so beautiful,” The Melty Way developer Gabriel Paquette told me. “Why not use it?”

It was a common refrain. Those I spoke to, most of whom were indie developers, disavowed AI, and many said they would never use the technology as it detracted from the human element of development. That’s perhaps not surprising, given that a recent GDC survey found that 52 percent of respondents think “generative AI is having a negative impact on the game industry,” which is up from 30 percent in 2025 and 18 percent in 2024. Some indie developers already go out of their way to show that their games are “AI free.” The largely negative reaction to Nvidia’s DLSS 5, which, in the publicly shown examples, added AI slop-like faces to recognizable game characters, almost certainly won’t make smaller developers more interested in the technology.

The general pitch for generative AI in gaming is that it might benefit both developers and players. In the most optimistic view of the technology, developers could use AI to help with tasks like debugging, QA, and idea generation, while players could use AI to help tailor games for themselves. Google Cloud executive Jack Buser, who helped launch Google Stadia and worked on PlayStation Now and PlayStation Home at Sony, says that generative AI is “the largest transformation in the games industry I have ever witnessed in my nearly 30-year career.”

But for many of those actually making games, the conversation is different. For instance, Adam Saltsman and Rebekah Saltsman, cofounders of the “collaborative” studio and publisher Finji, known for indie hits like Tunic and Chicory: A Colorful Tale, note that their works are defined in part by “a specific person or persons’ fingerprints.” In other words, a handmade, human quality, one that can include an element of surprise. ”You can show people what it is, but you are going to break all of their expectations when they go and play it,” Rebekah adds. That philosophy runs counter to the idea of utilizing generative AI in development. When I asked the Saltsmans if they would consider using generative AI for any of Finji’s games, it was a hard no. “Absolutely not,” Adam says.

Many developers told me that, in their view, AI-made games don’t look or feel like human-made games, at least right now. Audiences “don’t connect” with generative AI, according to Abby Howard, from Slay the Princess developer Black Tabby Games, adding that “I think it’s generic, I think it makes it feel cheap.” Rebekah is more blunt, saying that generative AI “just looks like crap.” For Matthew Jackson, who is working on the comedy game My Arms Are Longer Now, there’s another practical issue: “AI is so not funny.”

There are also legal problems that would complicate actually selling a game made with generative AI. Putting aside issues like the environmental impact of AI or concerns about the data AI is trained on, the Saltsmans tell The Verge they don’t think there is a legal framework to actually selling generative AI output. (This issue is also exacerbated by the fact that AI-generated art can’t be copyrighted.)

Finji isn’t the only publisher that isn’t accepting games made with generative AI. Panic, the publisher of Untitled Goose Game and creator of the Playdate, does not “have any interest in generative AI-created products,” cofounder Cabel Sasser tells The Verge. BigMode, the publishing company started by Jason Gastrow, aka videogamedunkey, requires developers to check a box with their application that says “I confirm that my game is human-made and does not include any use of generative AI.” Even Hasbro, which is now developing its own video games, isn’t using AI in its development pipelines, CEO Chris Cocks recently said on Decoder.

But perhaps what came up most often in my conversations at GDC is that using generative AI removes the craft from making video games. “The only way to get better at things is through the intense concentration of a career of applied craft,” Black Tabby Games’ Tony Howard-Arias says. Adam talked about how writing code can be “one of those things, like visual art, that pushes on your game design.” He points out that good programming is also good for players: “Things that are really hard to program are often really hard for a player to understand, too.” Alex Schleifer, cofounder of Ballgame developer Human Computer, says that the process of making games is just fun — and from that process, ”you’re also going to come to better ideas.”

There are concerns that AI tools might take away jobs from humans, which would both lower the pool of available positions in an industry already riddled with layoffs and provide new developers fewer ways to get their foot in the door. But despite the promised cost and efficiency savings — and that’s assuming an AI tool can even compare to what a human can do — this too would have problems. If you replace humans with AI, “where do you get new talent in the future?” Tony says.

Right now, the developers I spoke with believe crafting games by hand creates a more human connection. “We tell human stories,” Rebekah says. When you launch a game, there is a person that “you’ll never meet in your whole life that is playing a thing that you’ve spent thousands upon thousands of hours considering and working on.” Caring about their experience and that connection is “why we do this.”

Some indie developers I spoke with are open to the potential that generative AI in games could be useful for development or widely adopted down the line. The film and TV industry, for example, is seeing the rise of companies that build bespoke AI models to help with production, which could be a possible future for AI tools for game development. Maybe, at some point, AI will be more accepted, Paquette says. But for now, he prefers to do “100 percent” handcrafted work. “That’s something dear to me.”