On a stage in Korea, at Steam Indie Connect, Valve's business team faced developers asking an old question in a new form: what happens to creative work when machines can copy it faster than people can protect it. Valve has not yet found its answer, and it has chosen not to bar anyone from the tools that make the flood possible. Its position is that cloning has always existed and AI has only quickened it, so the real defense rests with creators who build something richer than a single copyable idea. Whether the platform can sort the thin from the sincere while keeping its doors open is still an
Valve Says It's Working on Steam's AI Slop Problem but Won't Restrict Developer Access
Valve won't act as gatekeeper over which tools developers use
So what did Valve actually commit to at this event?
Mostly a position. They won't restrict developer access or ban anyone for using generative AI tools. They said the focus is making Steam smarter.
Worth noting that no concrete new feature or policy was announced. It's a stated philosophy, not a fix.
Why are clones such a worry for indie developers?
Because early footage and demos can be used to build a cheap copy that reaches the market first. That makes showing your game early riskier.
The source says developers are growing more concerned. It gives no count of clones, so the scale is unmeasured here.
And Valve's answer to that?
Giardino said clones existed before AI and that AI accelerated them. He advised developers to keep adding features and personality rather than lean on one copyable mechanic.
That shifts the burden onto the original developer. Whether that's a satisfying answer is a fair question, and the reporting doesn't settle it.
What about the Next Fest complaints?
Wilkinson said recommendations are random for the first two days, then personalized. He said critical articles were based on that early window.
That's Valve's own account. We haven't seen data showing how things look after personalization starts.
Le Pouls
- Developers are hesitating to show their games early because generative AI lets copycats clone a project from footage or a demo before the original ever launches.
- Throwaway AI-made games are crowding Steam's storefront, and for small studios the struggle to be seen has become sharper than ever.
- Valve has refused to ban developers from using generative AI tools, arguing that doing so would make it a gatekeeper over creative methods.
- During Next Fest's opening days, recommendations are fully random by design, so weak games can be exposed before personalization sorts them.
- Valve admits the trade-off stings and says it is still searching for a fix, with no concrete new measure announced yet.
On a stage in Korea, at Steam Indie Connect, Valve's business team faced developers asking an old question in a new form: what happens to creative work when machines can copy it faster than people can protect it. Valve has not yet found its answer, and it has chosen not to bar anyone from the tools that make the flood possible. Its position is that cloning has always existed and AI has only quickened it, so the real defense rests with creators who build something richer than a single copyable idea. Whether the platform can sort the thin from the sincere while keeping its doors open is still an open chapter.
At Steam Indie Connect in Korea, Valve's Steam Business Team faced a room of game developers who wanted answers about the growing wave of generative AI games and AI-made knockoffs crowding the storefront. The honest reply was that Valve is still working out how to handle it.
The sharpest fear in the room was cloning. Developers have grown wary of showing their games early, because bad actors can use generative AI to copy a project from its footage or demo and get a poor imitation to market first.
Tom Giardino pushed back on the idea that AI invented this problem, noting that cloning predates the technology and AI has only sped it up. He ruled out the most direct fix. "We don't want to be gatekeepers telling developers which tools they should or shouldn't use," he said. The defense, he argued, belongs to the original creators: a single easily copied mechanic is a weak foundation, and games grounded in their makers' own perspective, personality and features are harder to imitate.
Discoverability was the other pressure point. Marc Wilkinson said the reports of a flood of low-quality games were based on Next Fest's first two days, when recommendations are completely random by design. Personalization arrives afterward, and players are then more likely to find games they actually like. The randomness exists so developers cannot learn to game the system, and Wilkinson acknowledged the cost: some low-quality games get exposed.
Giardino closed by admitting Valve does not consider its approach perfect. The stated goal is a smarter, more dynamic Steam rather than tighter limits on developers. No new measure was announced, so the next signal to watch is whether changes to the recommendation system follow.
At Steam Indie Connect in Korea, members of Valve's Steam Business Team stood on stage and took questions from an audience of game developers. One topic pressed hard: the wave of generative AI games and AI-made knockoffs showing up on the storefront. Valve's answer was that it is still working out how to handle it.
The clones are the sharper worry. Developers have grown more cautious about showing their games early, because bad actors are using generative AI to copy projects from early footage and demos. The copycats can then reach the market first with a poor imitation of the original.
Tom Giardino of Valve responded to the clone concerns by pointing out that cloning predates AI. In his telling, the technology has only sped the process up. He ruled out the most direct fix, a ban on developers who use these tools. "We don't want to be gatekeepers telling developers which tools they should or shouldn't use," he said, as relayed by InvaGlobal.
Giardino put the burden of defense on the original creators. Relying on a single mechanic that someone can easily copy is a weak position, he argued. Developers should keep growing their games with features, personality and creativity that come from their own perspective.
Discoverability was the other major thread. Developers in the audience said it is harder than ever to stand out among the throwaway AI games filling digital storefronts like Steam. The problem is especially acute during Next Fest, when developers are trying to find an audience before release.
Marc Wilkinson of Valve said the situation is not as bad as it looks. During the first two days of Next Fest, recommendations are completely random. Personalized recommendations begin afterward, drawing on user activity, and at that point players are more likely to find games they actually like. Wilkinson said the articles pointing to a flood of low-quality games were based on those first couple of days, before personalization began.
According to Valve, the randomness is deliberate. The algorithm has to start out random to keep developers from working out how to game it. Wilkinson acknowledged the cost: "We recognize that the trade-off is that some low-quality games get exposed." He added, "We are fully aware of the situation, and a significant number of people are thinking about how to solve this problem."
Giardino closed by conceding imperfection. Valve does not think it is perfect, he said, and there is plenty of room to improve. The starting point of its philosophy, in his words, is making Steam a smarter and more dynamic platform rather than limiting developer access. No specific new measure was announced, so what to watch for is whether changes to Steam's recommendation system follow.
Citations marquantes
We don't want to be gatekeepers telling developers which tools they should or shouldn't use.— Tom Giardino, Valve
We are fully aware of the situation, and a significant number of people are thinking about how to solve this problem.— Marc Wilkinson, Valve