Epic Games criticizes Valve for AI badges on Steam; Sweeney warns of harm to developers
Epic Games president Tim Sweeney recently expressed strong criticism against Valve’s guidelines for the use of artificial intelligence on the Steam platform. According to the executive, mandatory seals that identify games with AI are harmful to developers and could compromise the success of new titles on the market.
Sweeney’s statements were made during a meeting with PC Gamer, in Chicago, at the Unreal Fest event. Epic’s leader classified the rival company’s practices as harmful, reaffirming his support for the application of AI technologies in the game creation process.
How AI warnings become a stigma for games
Sweeney argues that Steam’s market power is a central factor in this issue. For a new game to gain visibility and be included on users’ wish lists, presence on the platform is essential, making the impact of the AI seal even more significant.
The executive equated the warning to the concept of “scarlet letter”, a mark of public dishonor imposed on a product. He noted that, when this warning appears, a part of the gaming community tends to organize to boycott the launch before it even reaches the public.
The origin of the expression goes back to the classic novel “The Scarlet Letter”, written by Nathaniel Hawthorne in 1850. In the plot, a character is forced to publicly display a red “A” on his clothes, symbolizing his conviction for adultery.
In this way, the “scarlet letter” became a symbol of visible stigmatization, a social punishment that constantly accompanies the individual. Sweeney uses this analogy to illustrate his perception: for him, the AI seal in games acts as a badge of disapproval, indicating to consumers something subject to repudiation.
A recent example mentioned by Sweeney was the game “Crazy Taxi”, which had a favorable initial presentation at a Microsoft event. However, the discovery of the AI warning in the title quickly changed public sentiment, which turned against the release.
Tim Sweeney considers Valve’s policy “irresponsible”
The criticism intensified when addressing Valve’s responsibility for implementing this standard. From the perspective of Epic’s founder, the company is imposing a very difficult decision on studios.
“I consider this attitude by Valve to be truly irresponsible,” said Sweeney. He added that “they shouldn’t act this way, as it immensely hampers the chances of success for any game developer.”
Sweeney’s line of thinking points to a dilemma: studios are forced to choose between giving up tools that increase productivity or using them and running the risk of facing public judgment due to the seal. According to him, rejecting automation represents a competitive disadvantage in relation to rivals that already employ it.
Details on what Steam actually asks of developers
The policy adopted by Valve, however, is more specific than Sweeney’s statements might indicate. In January, the Steam platform revised its guidelines, aiming to clarify what information about AI should be disclosed by developers.
Leia Também
Messi reaches 19 goals in FIFA World Cups with great free kick for Argentina in clash with Jordan
Cristiano Ronaldo shines beyond the pitch at the 2026 World Cup with business empire and family luxury
At 41, Cristiano Ronaldo reaffirms his historic legacy and prepares to shine in the 2026 World Cup
Current rules establish that development productivity improvements, such as the use of AI in code creation, do not require notification. The disclosure requirement applies exclusively to AI-generated content that is directly presented to the player, including visual, sound, textual elements and promotional materials.
The platform itself argues that, as this AI content does not go through the usual auditing methods, the responsibility for transparency lies with the developer. By July 2025, Steam registered 7,818 games with the AI seal, a significant growth of almost 800% compared to 2024.
How Epic Games defends the implementation of artificial intelligence
Sweeney argues that AI reduces repetitive tasks, allowing creators to focus on crucial aspects of games, such as narrative, setting and gameplay mechanics. He illustrated his vision with the metaphor of producing a vase of flowers.
He exemplified that manually modeling a highly detailed object, such as a flower vase, could reach an exorbitant cost of US$1 million (approximately R$5.2 million), in hyperbole. In his opinion, such an investment in a mere decorative item would be an unnecessary expense, since the true value of a game lies in the complete experience.
This perspective directly aligns with Epic Games’ future plans. The next version of its tool, Unreal Engine 6, will have integration with AI models such as Claude, from Anthropic, and Gemini, from Google, aiming to optimize various operations within the editing environment.
Sweeney himself admits that resistance to AI has its valid reasons. He acknowledged that in their early stages, some artificial intelligence companies engaged in questionable conduct, such as an episode in which a company was caught downloading terabytes of data from a torrent platform.
Contradictions at Epic Games: AI in focus after layoffs
Epic Games’ defense of automation occurs in a complex internal scenario for the company. More than a thousand employees were laid off this year, in a period marked by the decline in popularity of “Fortnite”.
The strong focus on artificial intelligence also had an impact on the creative aspect. Sjoerd de Jong, an experienced level designer, left the company after twelve years, and his departure was related to the new direction of Unreal Engine 6. Additionally, Poncle studio, creator of “Vampire Survivors”, revealed that it is reconsidering its collaboration with “Fortnite” after the release of information about Epic’s use of generative AI in creating game elements.
Critics also point out an inconsistency: while Sweeney advocates for properly licensed AI training bases, Epic’s artistic creation process uses tools like Nano Banana and GPT Image. These platforms, however, do not ensure strict control over the origin of the data used to train their models.
