Tim Sweeney Calls Steam AI Disclosures Irresponsible as Unreal Engine 6 Embraces AI

Epic Games chief executive officer Tim Sweeney has criticized Valve for requiring developers to disclose player facing generative AI content on Steam, arguing that the policy places a damaging label on games and makes it harder for studios to benefit from productivity tools. His comments arrive as Epic moves in the opposite direction with Unreal Engine 6, which is being designed around deeper integration with large language models, generative content tools, and AI assisted production workflows.

Speaking with PC Gamer, Sweeney said developers effectively need to release on Steam to reach its audience and benefit from features such as wishlists, but risk attracting hostility when an AI disclosure appears on their store page. He described the notice as a "Scarlet Letter" that can turn public discussion against a project before players evaluate the quality of the game or understand how the technology was used.

"I think it is really irresponsible of Valve. They should not do it, because it makes it much, much, much harder for a game developer to have a chance of success." Quote by: Tim Sweeney

Sweeney believes studios are being forced to choose between avoiding AI tools to protect their public image or using technology that could make production faster while accepting the risk of consumer backlash. He argued that this pressure is especially damaging for smaller teams attempting to compete with established live service games supported by hundreds of developers, years of accumulated content, and significantly larger production budgets. In his view, AI should reduce repetitive work around coding, asset production, testing, and world creation so artists, programmers, and designers can spend more time refining gameplay, narrative, environments, and the overall creative identity of a project.

However, Valve’s current policy is narrower than Sweeney’s criticism suggests. The official Steam Content Survey requires disclosure when AI generated material is included in content that ships with the game and is consumed by players. It also covers live content generated while a game is running, with developers required to explain the safeguards used to prevent illegal output. Valve clarified in January 2026 that internal efficiency tools, coding assistants, administrative work, and AI experiments that do not appear in the released product are not the focus of the disclosure system.

This means a studio can use AI to help identify a programming error, organize production information, generate internal tests, or assist with development tools without necessarily receiving a public Steam disclosure. The label becomes relevant when generated artwork, writing, voices, music, models, marketing material, or dynamic content reaches the customer. Valve says AI material is reviewed under the same legal and content standards applied to traditionally created assets, while acknowledging that creators have legitimate concerns about exploitation, consent, copyright, and misuse.

The disagreement therefore centers less on whether developers should use productivity tools and more on whether customers should be told when generated material becomes part of the product they purchase. Sweeney argues that disclosure creates stigma and discourages adoption, while supporters of Valve’s policy believe players have the right to understand how visible art, dialogue, audio, and other creative material were produced. The distinction is important because consumer objections often relate to training data, creative ownership, employment, and replacement of artists rather than every possible use of automated software.

Sweeney acknowledged that several early AI companies used questionable or potentially unlawful methods to obtain training material, but argued that licensed datasets and stronger industry practices could address those concerns over time. He also rejected the idea that games will soon be created entirely from simple prompts, stating that meaningful art still requires experienced creators to establish the vision, select concepts, shape assets, and build a coherent game.

"You are not going to create good art by giving a computer a prompt and having it spit out a mesh." Quote by: Tim Sweeney

Epic’s own plans demonstrate how central the technology has become to its development strategy. In its official roadmap for Unreal Engine 6, the company says large language models and generative AI will play a central role in helping developers produce content faster while retaining creative control. UE6 will include an open Model Context Protocol foundation with integrations for Claude, Gemini, and other models, allowing teams to connect their preferred AI services to engine functions and create custom workflows.

Epic is investigating AI for code indexing, automated testing, crash analysis, incident investigation, tool creation, and content authoring. Sweeney also described the ability to prompt an engine to create an initial city layout or a unique environmental object, which artists could then modify and refine instead of constructing every background detail manually.

The wider industry remains divided. Google Cloud has claimed that roughly 9 out of 10 surveyed developers already use some form of AI, although other surveys have produced lower adoption figures and the definition of AI can include everything from code assistance to final generated assets. At the opposite end of the debate, creators such as Hideo Kojima have rejected generative AI as an artistic tool, while remaining open to advanced technology that supports gameplay and interactive systems.

Worker concerns also remain central to the debate. Publishers increasingly promote automation as a way to control rising budgets, but the industry has simultaneously experienced extensive layoffs and studio closures. It would be inaccurate to attribute every employment reduction directly to AI, yet developers have valid reasons to question whether productivity gains will create more ambitious games, improve working conditions, reduce crunch, or simply allow companies to complete projects with fewer employees.

Sweeney is correct that a broad disclosure could become misleading if it treated a coding assistant, an internal spreadsheet tool, and a fully generated character performance as equivalent. Valve has already addressed part of that problem by limiting its policy to AI content that reaches players or appears in marketing. That distinction makes the Steam system more reasonable than the "Scarlet Letter" description suggests.

Transparency does not prevent developers from using AI, and disclosure does not automatically declare that a game is unethical or low quality. It gives customers information and leaves the purchasing decision with them. The solution should be clearer disclosures explaining what was generated, how it was used, and whether human creators remained responsible for the final work, rather than removing the information entirely.

Epic’s Unreal Engine 6 strategy may prove that AI can reduce repetitive production work without removing the human creative direction behind a game. However, Epic and other technology companies will need to demonstrate that their tools respect ownership, licensing, employment, and artistic consent. Greater efficiency is valuable, but trust will depend on what companies build with those savings and who ultimately benefits from them.

Should Steam continue requiring disclosures for AI generated content that players can see and hear, or do the labels unfairly damage games before launch?