Steam's AI labels aren't politics: they're basic respect for players

Since when is telling someone what's in the box a political act? That is the question buried under the latest round of Epic-versus-Valve trash talk, and a former Valve developer just asked it far more bluntly than Epic's boss would like.

Sweeney picked the fight, and a Valve veteran answered

Tim Sweeney has spent recent weeks hammering Steam. He called Valve's AI disclosure requirement "a political act" in a run of posts on X between October 23 and October 26, 2026, according to Gaming Route. In one of them he said that "forcing developers to write a political statement to be used against them by the hate mob they're helping organize is abhorrent." He also claimed Valve has "held gaming back for years" with Steamworks' social walled garden.

Chet Faliszek, who previously worked on Portal, Half-Life and Left 4 Dead, answered in a YouTube video, as GamesRadar reports. His verdict: "you're just treating players like they're stupid."

The "political" framing is the weakest part of Sweeney's case

Sweeney's argument leans on a hypothetical: should Valve force developers to list their political affiliation? Faliszek's reply is the right one: "Nobody's saying legitimately to do that." A label saying a game used generative AI is a statement about the product, closer to a frame-rate note or an age rating than a manifesto. We would go further: the slippery slope here is invented.

It also helps Faliszek that, as he points out, Epic's own Fab marketplace uses AI disclosures. By his account, Sweeney said he would keep them there even if Steam dropped its own because it "just makes sense" for business. That is a hard position to square with calling the same practice abhorrent. Either disclosure is a legitimate piece of product information, or it isn't. Epic seems to think it is when the shelf is its own.

Informed choice is not a mob

The real fear behind Sweeney's complaint looks like harassment: labels become ammunition. That is a legitimate worry, and we would not dismiss it. But the fix for a toxic audience is moderation and developer support, not hiding information from everyone else. Faliszek accepts that Sweeney doesn't care about AI usage, much as Faliszek himself doesn't care about a game's frame rate, yet he says he won't get mad that the information is listed because he understands some people want it. That is the sane middle ground: you don't have to share a concern to accept that customers can have it.

There's a market logic too. Players spend real money, and Faliszek's point is that Steam has a lot of people who want to buy games there. Giving them one more line of information costs a developer a few sentences. Withholding it asks buyers to trust blindly. Our view is that, with Valve business team members having owned up to Steam's growing AI slop problem, labels are not a luxury but a basic filter.

Where Faliszek is less convincing, and what to do

To be fair, this is a former employee defending his old employer, and some of his rhetoric ("AI psychosis," "crashing out") is more roast than argument. His strongest card is the challenge to put Alan Wake 2 on Steam if Epic truly opposes walled gardens. Whether Sweeney answers remains to be seen. On tooling he even concedes Epic has "an amazing set of tools," including a backend he says surpasses Valve's. Fair enough, and it shows the dispute isn't a simple good-versus-evil story.

But on the narrow question, we side with the labels. Developers who are proud of how they use AI should have no trouble saying so; players who don't care can scroll past. The only people who lose from disclosure are those hoping nobody asks. Our recommendation: treat the label as information, read it, and judge the game. Stores should keep it, and Epic, which already does it on Fab, should stop pretending it's a loyalty test.