Valve's Steam Next Fest Response Leaves AI Clone Risks Open

Valve has acknowledged that Steam Next Fest’s early visibility system exposes players to some low-quality games. At a developer event in South Korea, business development representative Marc Wilkinson said games appear on a largely random basis for the festival’s first two days before personalized recommendations begin. Valve knows the trade-off and has people considering how to address it, he said. His remarks reached English-language outlets through an InvenGlobal account translated from Korean, so the wording of quoted remarks should be read in that light.

The exchange followed complaints about games that players and developers characterized as “AI slop.” A developer who had participated in the latest Next Fest also described poor engagement and raised complaints about low-quality games appearing prominently, Video Games Chronicle reported. The complaint puts two questions on Valve’s desk: how a game gets seen, and what happens when another developer can quickly imitate it. The same discovery system touches both, but it cannot answer both on its own.

What the First Two Days Are For

Wilkinson gave a reason for the randomized opening. It distributes early exposure among developers and limits the advantage someone might gain by entering early to secure immediate visibility. After that period, recommendations reflect individual players’ interests. The design gives an unfamiliar game a chance to appear before the platform has enough engagement with it to favor or bury it.

That is a defensible choice for a festival where developers compete for attention. A filter that favors established interest from the opening hour could reinforce the advantage of games that already have an audience. Random exposure avoids making early visibility depend entirely on that prior attention. It also means a player’s first encounters can include work they consider poor, while a developer with a strong demo cannot assume the opening period will deliver an interested audience. Wilkinson acknowledged the former cost; the latter follows from a system that waits before personalizing recommendations.

Staff have thought about the visibility problem, with Wilkinson acknowledging the low-quality-game trade-off. That is a statement of attention, rather than a timetable or a proposed replacement for the current design. Players asking for better recommendations and developers asking for a fair chance are pressing on different parts of the same mechanism. Tightening the first two days around demonstrated interest might help some players find appealing demos sooner, while making it harder for games without existing attention to break through. Keeping the opening broad protects that opportunity, but asks players to sort through more games themselves.

The phrase “AI slop” can conceal that trade-off. A player can dislike a game’s presentation without knowing how its assets or code were made. A developer using an AI tool is a separate question from whether the resulting demo appeals to anyone, and both are separate from a threat to copy somebody else’s project. Treating all three as a single moderation category would make any policy sound simpler than the job Valve actually faces.

A Different Problem for the Developer Being Copied

At the event, a developer raised alleged emails threatening to clone games with AI and release them first unless their creators paid. Valve business team member Tom Giardino said he had not previously heard of that situation. He indicated Valve would look into reported cases, while describing extortion as principally a matter for law enforcement. The email claims came through a developer’s question; they do not establish how common such threats are or whether the threatened games were cloned.

Giardino also said Valve did not want to tell developers which tools they could use. He described Steam’s general preference as giving weight and discoverability to originals rather than removing copies and clones. His advice to creators was to keep adding features and personality that a copycat could not easily reproduce, and to build a community that recognizes and prefers the original. The remarks passed through InvenGlobal’s translated account.

The strongest case for that approach starts with an operational problem: banning a tool does not identify every imitation, and a broad rule against games that resemble other games would require Valve to make difficult judgments about creative work. Giving an original better placement could direct players toward it without closing the storefront to legitimate developers. Community loyalty and continued development might also help a studio once players know which team made the game.

The sequence is harder for a small developer who has yet to find those players. Building trust requires contact with an audience. Adding features requires development time. A threat to release a lookalike first creates pressure before either asset has fully formed. Valve controls the recommendation and storefront systems that could distinguish an original, while its advice asks the individual creator to supply the continuing work and the audience relationship. That division does not prove Steam caused any cloning, and better recommendations alone would not stop an extortion attempt. It identifies what each proposed remedy can reach.

Digital Sun Games offers a concrete measure of the pressure. The studio behind Moonlighter is delaying a public playtest of its next game, ReVamp, because it fears AI clones could reach release ahead of it, Game Developer reported. A delayed playtest is the studio’s decision about when to show work, whereas Steam’s admission and discovery policies govern what happens on the platform. Together, they show the timing problem: a developer may want players to see a game early, yet fear that showing it gives an imitator an early look too. Advice to grow a community is harder to follow when the public demonstration that could help grow it is itself being delayed.

The Old Access Argument Meets a Faster Imitation Worry

Valve has faced versions of the filtering request before. Giardino recalled developers asking to filter out titles when Steam Greenlight appeared, then making similar requests around the self-publishing system Steam Direct. He presented Valve’s longstanding preference as improving discovery so stronger games rise rather than restricting developer access.

In 2018, Valve’s Erik Johnson put its broad storefront position plainly: allow games onto Steam except those the company considered illegal or “straight up trolling,” as quoted by Eurogamer. That history explains why Giardino’s response favors placement and developer choice over a new exclusion rule. A broad-access policy can benefit an unknown original as well as someone trying to imitate it. The present worry adds a deadline to that old bargain: a creator concerned about a rapid copy may need protection before an audience has had time to identify the original.

Those are different decisions for Valve to make. It can reconsider how Next Fest allocates early impressions without banning AI tools. It can consider how an original is recognized without treating every similar game as misconduct. And it can investigate a reported threat while acknowledging that a storefront operator is not a law-enforcement agency. Each choice has a cost or a limit. For the developer deciding whether to put a demo in public, the practical question is whether Steam can make the original findable soon enough for the community Valve recommends to form.