Epic CEO Slams Steam for Making Studios Disclose AI Use

The debate over how games should disclose their use of artificial intelligence has a new flashpoint, and it pits two of the biggest names in PC gaming against each other. Epic Games CEO Tim Sweeney has publicly criticized rival Valve for requiring studios to disclose when they use AI in game development, calling the policy "really irresponsible" and likening the resulting label to a "Scarlet Letter" that turns players against a game before it even launches. It's a pointed accusation, and one that deserves a closer look.

Since 2024, Valve has required developers to state on Steam whether generative AI was used in creating their games, with the information appearing on each store page in an 'AI Generated Content Disclosure' section. According to reports, the policy was tweaked in January 2026 to clarify that the two types of AI requiring disclosure are AI used to generate content for the game and AI content generated live during gameplay, while "AI-powered tools" such as coding assistants don't need to be flagged at all.

Sweeney's 'Scarlet Letter' argument

In an interview with PC Gamer, Sweeney argued that any developer who declares they used AI to generate content is effectively painting a target on their own back. His concern is that Steam's reach makes the disclosure unavoidable for anyone hoping to succeed.

If you want to launch a game, and get it as widely publicized as possible, you've got to put it on Steam so people can wish list it, and if you want to play it on Steam, then you have to get this Scarlet Letter of AI attached to your product, and now there is a hater community trying to kill the game.

— Tim Sweeney, Epic Games CEO

Sweeney went further, framing the rule as something that actively harms developers' odds of survival. "I think it's really irresponsible of Valve," he said. "They shouldn't do it, because it makes it much, much, much harder for a game developer to have a chance of success. You have to choose from either not using tools that can make you way more productive, and probably failing due to competition that does." In his telling, the policy forces studios into a no-win choice between productivity and public goodwill.

Epic's own AI ambitions

That argument lands differently given where Sweeney is standing. His comments arrive just as Epic has begun talking up Unreal Engine 6 and its heavy emphasis on integrating AI models like Claude and Gemini directly into the development workflow. In other words, Epic has a clear stake in normalizing AI-assisted creation: a context critics will note, and supporters will say simply makes him an informed voice on where the industry is heading.

At a State of Unreal event in Chicago earlier this month, Epic demonstrated how large language models can be used directly inside Unreal Engine 6 to generate content. In one demo, a Claude prompt window furnished a virtual apartment on request, pulling items from the asset library; in another, the team changed a city scene's lighting simply by asking Claude to alter the time of day, or by feeding it a static photo as reference. Epic stressed that developers retain final control and can manually adjust any result, a detail clearly aimed at the trust questions these tools raise.

None of this is new territory for Sweeney. Last year he argued on X that it "makes no sense for game stores" to flag AI use, since AI "will be involved in nearly all future production." That's the crux of his position: if AI becomes ubiquitous, a disclosure label stops being meaningful information and starts being a stigma.

The case for disclosure

There's a counterargument that Sweeney's framing leaves out, and it's why Valve's policy exists in the first place. For many players, knowing whether a game's art, writing, or voices were generated by AI is exactly the kind of information they want before spending money: a matter of consumer transparency, not mob justice. Supporters of disclosure also point to the human side: generative AI trained on existing creative work touches directly on the livelihoods of artists, writers, and voice actors, and a label lets buyers factor that into their choices. From this view, the 'hater community' Sweeney describes is, at least in part, simply customers exercising informed preference.

The nuance Sweeney glosses over is that Valve's January update already tried to draw those lines, exempting development tools from disclosure entirely and focusing the requirement on AI-generated content that ends up in the shipped product. A coding assistant, an AI-upscaled texture, and a fully machine-generated character are not the same thing, and the policy doesn't treat them as if they are. Dismissing it as a broad brush misreads what it actually requires. The backlash Sweeney describes is real, but it tends to concentrate around games that leaned heavily on generated assets, not tools that helped a small team ship something they couldn't have otherwise.

For what it's worth, this is one case where the scales tip in a pretty clear direction. Valve has spent the last decade building a track record that most of its competitors simply can't match: a refund policy that other storefronts had to scramble to copy, the Steam Deck proving consumer hardware can be done right, and a general reluctance to pull the kind of platform shenanigans that have become standard practice elsewhere. Epic, for all Sweeney's genuine passion for developer-friendly ecosystems, spent years funding a storefront expansion through timed exclusivity deals that most players never asked for. When Valve decides something is worth requiring, that decision carries weight. Believing it must be driven by bad motives takes more effort than the evidence supports. Source: PC Gamer's interview with Tim Sweeney.