Epic boss Tim Sweeney blasts Steam for putting AI tags on games — says move is ‘irresponsible of Valve’

Epic Games Store
(Image credit: Epic Games)

After Epic revealed its plan for Unreal Engine 6, CEO Tim Sweeney said in an interview that Valve should stop requiring developers from disclosing if they used AI in making their game. The Epic chief executive made the statement to PC Gamer as he was talking about the use of AI tools in game development. The upcoming version of Unreal Engine 6 will come with AI integrations, which will supposedly make it easier for programmers, developers, and everyone else working in the gaming industry to build games. However, the use of AI tools still carries a stigma among players to the point that titles with this disclosure get significantly fewer reviews and are often viewed less favorably.

He says that AI tools are useful for streamlining boring, repetitive, and menial labor, like reviewing code for over an hour to find an error or doing the rigging work required to make a 3D model move realistically. While it was unfortunate that some AI tech companies trained their models on stolen data, the Epic chief executive said that the AI industry has changed and is now moving towards better practices, especially when it comes to training data. He even pointed at Adobe, which he says is ensuring the provenance of the data it used for training its AI. However, PC Gamer pointed out that Epic uses Nano Banana and GPT Image, which do not claim stringent AI training data controls, and that the Unreal Engine’s AI integrations include models such as Gemini, which have been accused of copyright infringement.

“I think the main usage case that we've seen within Epic, and we're seeing developers actually find gainful, is using AI to reduce the drudge work. The software is still architected by software architects, and they're still writing the important parts of code, and artists are still coming up with a creative vision for characters, deciding between concepts,” Sweeney told the publication. He also added, “If you want to launch a game, and get it as widely publicized as possible, you've got to put it on Steam so people can wish list it, and if you want to play it on Steam, then you have to get this Scarlet Letter of AI attached to your product, and now there is a hater community trying to kill the game. I think it's really irresponsible of Valve. They shouldn't do it, because it makes it much, much, much harder for a game developer to have a chance of success.”

Latest Videos From

The Epic executive has a point here, as AI can indeed be a useful productivity tool if used correctly and ethically. But even if that was the case, Valve still requires developers to add the AI-generated content disclosure, which could have a negative impact on reviews and ratings. While the Steam platform does not require the disclosure of the use of AI-powered tools in the developers’ workflow, they’re required to note it if AI-generated assets appear in the game or in marketing material. This would presumably include assets that have been partially built using AI tools or were based on AI-generated content.

According to market research platform Game Oracle, titles that had the generative AI disclosure received 53% less reviews than the same type of game that did not have it. Furthermore, it’s more likely to get a negative review. However, it also conceded that several other factors could be at play here, like studios substituting creativity and the proper development process with AI tools, resulting in titles that are obviously AI slop.

Sweeney argues that AI is just a tool for productivity that could help game developers create unique content. The AI disclosure warning on Steam games, in theory, should not influence how games are perceived — after all, what should matter to gamers is the experience of playing the game. Unfortunately, the controversies surrounding AI are negatively affecting the reputation of titles that use these tools. When paired with studios that substitute AI for real creativity, it’s understandable why this AI warning has such a negative impact on game titles.

Google Preferred Source

Follow Tom's Hardware on Google News, or add us as a preferred source, to get our latest news, analysis, & reviews in your feeds.

Jowi Morales
Contributing Writer

Jowi Morales is a tech enthusiast with years of experience working in the industry. He’s been writing with several tech publications since 2021, where he’s been interested in tech hardware and consumer electronics.

  • ezst036
    Listening to the whines of Tim the Corporatist turns my ear into tin. This guy is a complete nothing burger. He charges more for games on his store than Steam does on their store, he doesn't innovate jack, he just waits for other people to come up with new stuff and then plays to popularity and tries to make the most buck for the least amount of money.

    I can't stand corporatists. He is a billionaire and its not enough billions, he is a billionaire and he still keeps whining. Hey Tim, go lower your insane prices and we'll listen to you.
    Reply
  • hotaru251
    90% of the "ai" games are slop cash grabs.

    if games w/ "ai" start being good then maybe the tag will stop being a bad thing..but until then clearly stating if uses ai is correct path.
    Reply
  • -Fran-
    So disclosing what tooling you use for the game is a bad thing? What is that backwards logic all about?

    What a friggen' clown.

    Regards.
    Reply
  • usertests
    Sweeney smells blood in the water, and must exude cringe immediately.
    Reply
  • TerryLaze
    Boss of the company Epic != Epic boss

    Epic's boss
    https://platform.theverge.com/wp-content/uploads/sites/2/chorus/uploads/chorus_asset/file/25157482/236883_Epic_Vs_Google_10_CVirginia.jpg?quality=90&strip=all&crop=0%2C0%2C100%2C100&w=1200EPIC BOSS
    https://static0.thegamerimages.com/wordpress/wp-content/uploads/2017/02/bowsermario.jpg
    Reply
  • Klathra
    ezst036 said:
    Listening to the whines of Tim the Corporatist turns my ear into tin. This guy is a complete nothing burger. He charges more for games on his store than Steam does on their store, he doesn't innovate jack, he just waits for other people to come up with new stuff and then plays to popularity and tries to make the most buck for the least amount of money.

    I can't stand corporatists. He is a billionaire and its not enough billions, he is a billionaire and he still keeps whining. Hey Tim, go lower your insane prices and we'll listen to you.
    The developers set their prices, not the platforms. Steam takes a much larger cut of the sales.
    Reply
  • timsSOFTWARE
    I think the biggest issue with AI tags is that they are somewhat open to interpretation and rely on honesty in self-reporting. There is a line to be drawn somewhere between, "I did a Google search once, and Gemini suggested something to me and it helped me with development", and "All of the assets and code in the game are AI-generated - I actually can't code or create art myself." But without that line being clearly defined, it's left open to interpretation - and honesty.

    Ie., if none of the shipping code and content in the game are AI generated, but you used AI concept art, AI placeholder assets, AI advice and sample code - is that an AI-developed game or not? And if you have to count that as AI development, then how far do you have to back off from there, before it can be counted as a non-AI production.
    Reply
  • alan.campbell99
    Using LLMs to review code, AI that will inevitably hallucinate at some point , probably won't guarantee completely reliable product. How often do software engineers have to check outputs?
    I mean, on the face of it removing that sort of drudge work sounds appealing but if it ends up that a human still has to review the reviewer will it work out well in the end?
    Reply
  • Arkitekt78
    Hes allowed to be wrong. AI games are crap, and they should be flagged. Doesn't keep people from buying them, but gives the consumers all the information so they can avoid garbage slop games if they want to.
    Reply
  • derekullo
    alan.campbell99 said:
    Using LLMs to review code, AI that will inevitably hallucinate at some point , probably won't guarantee completely reliable product. How often do software engineers have to check outputs?
    I mean, on the face of it removing that sort of drudge work sounds appealing but if it ends up that a human still has to review the reviewer will it work out well in the end?
    Humans hallucinate as well.
    That's why we have peer review.
    Reply