Picture a solo developer. One person, a laptop, a game they built over two years in their spare time. They used an AI tool to help write some placeholder dialogue. Now Steam makes them stamp an “AI Generated Content Disclosure” on their store page, and buyers treat it like a health warning on a cigarette pack.

That is the fight happening right now, and most people are cheering for the wrong side.

Here is what Valve actually did. Since 2024, Steam requires developers to disclose whether generative AI made any content that ships in the game. Art, text, audio, anything players see. In early 2026 Valve tightened the form. Some commentators started calling the label a “scarlet letter.” Epic CEO Tim Sweeney called Valve’s approach “really irresponsible” and said Epic would not require the same labels on its store.

My argument: the label is fair, but the way buyers read it is not, and that gap punishes small developers most.

A big studio can absorb a stigmatized label. Ubisoft ships a game, the label sits there, and the marketing budget drowns it out. A one-person team has no budget. The label is the first thing a browsing buyer sees, and a chunk of them will click away on reflex. The transparency that was supposed to protect players ends up sorting games by fear instead of quality.

Valve is not wrong to require disclosure. In one case Valve told a developer it could not ship a game using AI art trained on data with unclear copyright, unless the developer proved they owned the rights. That is copyright law, not an anti-AI crusade. Buyers deserve to know what they are buying.

The counterargument is real: labels stop players getting scammed by low-effort shovelware churned out in a weekend. True. But Steam was already drowning in shovelware before generative AI existed. The label does not fix that. It just tags a category and lets buyers’ bias do the rest.

If we want honest labels, we have to read them honestly. A disclosure is information, not a verdict. Judge the game, not the sticker.