Valve, the gaming behemoth known among other things for operating popular gaming platform Steam and not releasing Half-Life 3, has taken significant steps in allowing games that integrate Artificial Intelligence (AI) content on its platform. We see no way this could be a catastrophe. (/s)
This move comes after Valve previously declared that Steam would not accept games utilizing AI. Valve has since clarified its stance, indicating that it was actively working on incorporating AI into its existing review policies, aligning them with current copyright and trademark laws.
The announcement of this change was posted on the Steamworks Development news page. Valve officially unveiled changes to how it reviews games incorporating AI technology for availability on the Steam platform.
These modifications aim to facilitate the release of what Valve claims is a “vast majority of games” that are newly developed utilizing AI. The key alterations include updates to the Content Survey, a form that developers complete when submitting a game to Steam.
The survey will now include a section where developers must disclose whether AI was utilized during development, specifying whether it was pre-generated (for creating assets pre-release) or live-generated (for generating content during gameplay).
For pre-generated content, developers must assure Valve that their game will not include illegal or infringing content:
Any kind of content (art/code/sound/etc) created with the help of AI tools during development. Under the Steam Distribution Agreement, you promise Valve that your game will not include illegal or infringing content, and that your game will be consistent with your marketing materials.
On the other hand, for live-generated content, developers are required to outline the guardrails implemented in their AI to prevent the generation of illegal content. Additionally, Valve has introduced a new system on Steam that empowers players to report illegal content within games featuring live-generated AI content.
Through the in-game overlay, players can easily submit reports for content that they believe should have been identified by appropriate guardrails on AI generation. This is a nice silver lining, and hopefully players will be vigilant enough to use it.
It should be noted that the restriction on “Adult Only Sexual Content that is created with Live-Generated AI” remains in place, indicating a limitation on this type of content release. This is likely due to reports that datasets such as LAION-5B include illegal CSAM and CP content.
Live-generated AI with this data would be more difficult to determine and filter out. This is a crucial step to avoid controversies similar to the one faced by AI Dungeon in 2021, which drew criticism for generating inappropriate content featuring children.
This new policy is similar to Steam’s rival, the Epic Games store, and its CEO Tim Sweeney’s stance. Sweeney explicitly welcomes games powered by AI, including those using ChatGPT, on the Epic Games Store. In a response on X.com (formerly Twitter), Sweeney emphasized that Epic Games does not ban games for “using new technologies.
Put it on the Epic Games Store. We don’t ban games for using new technologies.
— Tim Sweeney (@TimSweeneyEpic) September 2, 2023
However, some developers continue to resist the idea of AI replacing the role of skilled specialists in game development. One example is Warframe’s Digital Extremes Creative Director Rebecca Ford, who expressed in a CBC interview her strong stance against the idea of anything creative taking the place of skilled specialists.
This echoes the sentiment of the majority of artists, designers, programmers, and other creators, as current AI allegedly infringes on data without compensation or credit. Major lawsuits are already ongoing in relation to these issues.
This new move by Valve reflects the increasingly volatile landscape of AI integration in the gaming industry. Attempts are being made to use innovation as a rationale for sidestepping the ethical and legal implications.
These policy changes on the Steam platform could provide a framework to navigate the responsible use of AI in game development. On the other hand, it could also lead to decreasing quality and an avalanche of mediocre to terrible AI-generated games – in which case we’ll just need to vote with our wallets.
