Epic Games CEO Tim Sweeney has once again challenged Steam's requirements for developers to disclose generative AI use in games, describing the policy as "abhorrent" in a new social-media response. His latest comments arrive during an especially active period for Epic's public embrace of AI-assisted game development, while players and creators continue to debate whether storefront transparency is essential or unfairly stigmatizes developers.
Steam requires developers to provide information about how generative AI was used in a product submitted to the PC storefront. The disclosure can cover material created before release as well as AI-generated content that could appear during a game's live operation. In practice, descriptions vary widely: one recent example for Call of Duty: Modern Warfare 4 states that the development team uses generative AI tools to assist with some in-game assets.
Sweeney objects to the underlying principle of requiring that statement at all. Replying to Matt Workman, who argued that an AI-generated-content disclosure is absurd in 2026, Sweeney said that forcing developers to write a political statement that can be wielded against them by a "hate mob" is abhorrent. The remarks continue a long-running campaign by the Epic executive against Valve's approach.
A continuing disagreement over transparency
Sweeney had already criticized Steam's AI disclosures by November 2025, arguing that they made little sense in a future where AI would be involved in nearly all production. He returned to the subject in June, likening labels of this kind to a "Scarlet Letter" and arguing that they could make it far more difficult for a developer to find commercial success.
For useful background on this topic, read Call of Duty: Modern Warfare 4 review.
That view reflects a broad prediction among many technology advocates: AI tools, they argue, will become as ordinary in production pipelines as image-editing software, game engines, version control, or other forms of automated assistance. From that perspective, a mandatory disclosure risks treating a broad category of tools as inherently suspect, even when their role in a game may be limited.
Steam's policy, however, is not framed as an outright ban on generative AI. It is intended to give customers information about where and how AI systems shaped a game, while also asking developers to account for content that could be generated during play. Valve has said that the field is changing quickly and acknowledged both the concerns of artists and creators over exploitation and misuse and the possibility that developers may find ethical ways to use AI in entertainment.
The disagreement is therefore not simply over whether games can use AI. It is over who should be informed, what level of detail is appropriate, and whether revealing that use creates necessary consumer clarity or an unjust commercial penalty.
Epic is making its AI ambitions more visible
The timing of Sweeney's latest comments is notable because Epic has recently been unusually direct about its own experimentation. The company said it "vibe coded" the Epic Games Launcher as part of work on its redesigned PC storefront, using a phrase associated with directing AI coding tools through natural-language prompts rather than manually writing every portion of software.
Epic also announced an Unreal Engine game jam that initially required AI use and advertised unlimited agentic support for participants. A game jam is typically a short, intensive event in which creators build playable projects under tight time constraints. Following pushback, Epic revised the event's framing and said participants could use as little or as much AI as they wanted.
That adjustment illustrates the difficulty companies face when they present generative AI as a creative opportunity. Even developers open to technical experimentation may object to mandatory use, particularly in a community event where participants have different artistic values, skill sets, accessibility needs, and concerns about rights. Optional access may broaden the range of ways creators can participate, but it does not resolve the wider disagreement over what AI's role should be.
Epic's stance is clear: the company sees AI as a meaningful part of the future of game creation and software development. Sweeney's criticism of disclosure rules appears inseparable from that larger belief. If AI assistance becomes commonplace, he argues in effect, a storefront label could eventually identify a huge portion of the market rather than a narrow and exceptional type of product.
Why players and creators want labels
For critics of generative AI, the case for disclosure begins with a straightforward question: if the technology is beneficial, why should developers be reluctant to identify its use? A second question follows quickly. If audiences react negatively to a label, what specifically is driving that reaction?
The answers are not limited to one issue. Legal uncertainty is a central concern. Generative models have been trained on vast bodies of material, and disputes over copyright, consent, licensing, and ownership remain active across the broader technology and creative sectors. Game developers may face different levels of risk depending on the tools they use, the assets they create, the data behind a model, and the terms offered by the provider.
Creative concerns are equally important. Artists, writers, musicians, voice performers, designers, and other specialists have argued that generative systems can devalue skilled work or encourage companies to replace essential craft processes with cheaper automated output. The concern is not only that a final image, line of dialogue, or piece of music could be weak. It is that production decisions may shift away from human creators whose judgment, experience, and collaboration are fundamental to the medium.
There are ethical and environmental questions as well. Critics point to the energy and computing demands associated with major AI systems, alongside persistent arguments that model makers have not adequately addressed the use of material gathered without creators' permission. These issues are especially charged in games, an industry built on art, music, storytelling, performance, and communities that often closely follow the people behind the work.
Transparency labels give consumers a way to make their own decisions amid that uncertainty. A player may be comfortable with AI used for internal programming assistance while refusing to buy a game that uses generated art, voice work, or music. Another may not care about the distinction. Steam's disclosure system allows those preferences to be expressed at the point of purchase rather than requiring players to investigate a game's production history on their own.
A crowded storefront makes the issue more immediate
The debate has become more urgent as the volume of AI-aided releases and demos rises. During a recent Steam Next Fest showcase, users identified more games using AI assistance than in prior events, alongside a total demo count that was nearly double the volume seen in June 2026. The larger flow of games has renewed worries that digital storefronts could become harder to navigate as cheaply produced, rapidly assembled projects compete for visibility.
Recommendation systems can reduce the reach of low-interest games after player behavior provides enough signals, but that does not eliminate the discovery problem. Before those systems have meaningful data, an influx of releases can make browsing more difficult for customers and can make it harder for smaller teams to attract attention. Labels are one possible sorting mechanism, whether players use them to avoid AI-produced content or simply to understand a project's development methods.
There is also concern over rapid cloning. Developers have increasingly described situations in which popular concepts are copied at remarkable speed, and some have even cancelled planned demos because they fear their work will be replicated before a full release. Generative tools are not the only reason imitation happens in games, but their ability to accelerate asset creation, prototyping, and software work has intensified anxieties around copycat releases.
Not every use case draws the same response
Public opinion on AI in games is not uniform. Resistance is often strongest around generated artwork, music, writing, voice performances, and other material that audiences associate directly with human expression. AI assistance for programming is sometimes received more favorably, particularly when it is viewed as an advanced productivity tool rather than a replacement for creative labor.
Even that distinction is not absolute. Some developers welcome code assistance while remaining wary of errors, security issues, unclear licensing, and the effect on learning and hiring. Others oppose AI across all stages of production. Acceptance also varies by region and market, with AI generally seeing more support in some Asian game markets than in many Western communities, even though opposition exists worldwide.
Valve co-founder Gabe Newell has expressed a particularly expansive view of machine-assisted programming, while Valve's own developers have used AI in some contexts. That does not remove the company's disclosure requirements, which suggests Valve sees room for both experimentation and customer-facing transparency.
Sweeney's "hate mob" characterization is likely to further sharpen an already divisive discussion. To supporters of disclosure, the phrase risks grouping together consumers, artists, and developers with legitimate objections under a dismissive label. To opponents of the policy, it captures the fear that any AI-related admission can trigger an outsized backlash regardless of the actual scope or purpose of the tool's use.
Neither side appears ready to retreat. Epic is continuing to champion AI-enabled workflows, while Steam is retaining a policy built around disclosure rather than prohibition. As AI tools become more embedded in game production, the pressure will be on storefronts, studios, and platform holders to draw clearer lines between experimentation, accountability, consumer choice, and the creative work players expect to support.
Community
Discussion
Start the conversation.
No comments have been posted yet.