I'm guessing from Valve's point of view the reluctance to allow this has to do with potential rights issues surrounding chat gpt and related technologies. It's bad that their communication is so opaque and I don't like their near monopoly status but adding this feature, even if it was semi hidden away, was not a good idea. Hopefully the op on reddit is able to remove it and resubmit it.
The problem is the op did resubmit the game without AI, but was still rejected for the same reason:
I explained exactly how this worked and that there was no AI-content directly in the build, and even since then issued a new build without this mod ability just to be super safe. [...] Then, today, I randomly received an email that my app has been retired with a generic 'your game contains AI' response.
Interestingly the offer to train a model with only open material to avoid copyright issues wasn’t sufficient. It seems to be a general “AI bad, no AI for you” unthinking knee jerk.
I’m worried Valve will strangle what I think is the most exciting use of AI - artist trained models generating NPC content, lore, and procedural game content to create rogue like RPG of arbitrary complexity.
I doubt it's unthinking. Valve hews pretty Libertarian, historically they've been very reluctant to ban even egregious games. If they're disallowing something it's probably because they don't feel they have any other option (presumably because they feel they'll expose themselves to legal liability and jeopardize quality of the platform).
My guess would be that they can't verify what AI you used, at least without investing the kind of resources required to actually make case by case decisions.
I'm not saying this is your position, this is somewhat tangential, but I am confused why I seem to hear from people that both AI is a force of nature and companies stand in it's way at their peril, but also that it's a delicate flower and that any kind of management may strangle it in the crib. To be clear I know you aren't making both claims, I don't want to put words in your mouth, but it does feel like it's coming from the same general community of people. And I'm perplexed by that.
But yeah, I would also be interested to play and perhaps developing games like that. If someone is developing a game like that and they wants to hire a backend dev with experience integrating with models in a realtime context, shoot me an email. (I'm half kidding, I don't expect someone like that to happen to be in this particular thread, but only half.)
Yes I’m sure their original rationale is thought out, but reading the Reddit threads attached of the various developers trying to navigate the policy doesn’t give faith that there’s thought going on in application.
How do they verify copyright material isn’t in static assets? I can’t imagine it’s very thorough. I imagine most of it is based on attestation.
Regarding your question, I don’t share those views but can understand how they exist simultaneously. With an open hand, like was afforded internet and technologies, it could shape the world in hard to imagine ways as did the internet. However like the early internet it could have been strangled in the crib through knee jerk regulation and prohibitions. I think AI model builders and hosters want safe harbor similar to content hosters and providers allowing the end use of the technology to be regulated but the providers to be allowed to operate on a best effort basis. For instance the copyright question here is strangling any game that uses a generative model. A reasonable legal outcome is that it’s fair use to incorporate copyright material, but the output must not violate copyrights.
If Valve had no good way to verify that the model is trained on open data then allowing those models is the same as allowing all AI. Doesn’t seem like a knee jerk.
I’m dubious they need more than an affidavit from the publisher and developer attesting their model was trained without copyright material. The publisher and developer are the ones directly liable for copyright infringement. Valve does it’s own verification of game content and that combined with a legal declaration from the publisher should be enough to insulate valve.
I’d note they also have no way to reasonably unpack all game assets, visual, text, and audio, and verify there’s no copyright material. To my memory there are several cases of video game publishers including copyright violating assets. It seems like a copout to take on faith a no copyright violation attestation for static content but not on models.
Steam is a publisher, no? (ETA: I guess they're not necessarily the publisher. From Wikipedia: "Initially, Valve was required to be the publisher for these games since they had sole access to Steam's database and engine, but with the introduction of the Steamworks SDK in May 2008, anyone could integrate Steam into their game without Valve's direct involvement.").
I'm sure you're right that's all they would need to win a lawsuit, but they would probably prefer to avoid one altogether. Or at least not to be the first publisher sued.
There's no reason to suspect a generic asset is infringing on it's face, but for models trained on the open internet it's an entirely unanswered question.
Except in some of the linked Reddit threads the game authors explicitly said they would train models with curated content that’s under their control and open in license, which is not the open internet.
Indiscriminate crawler trained models are one thing, but models trained by the game developer directly should be handled differently. In that case they actually are the same as a generic asset, as they were assembled and curated by the game developer with an understanding of copyright compliance. But there are cases where they didn’t - particularly with audio.
I agree that morally and presumably legally these are completely distinct. But from a policy perspective, they aren't going to be able to verify that the output came from one model or another. So allowing any models is tantamount to allowing all models, because people who wanted to use eg Midjourney could simply lie.
If you tasked a human being with sufficient knowledge, latitude, and resources to investigate a game with a significant fraction of their time (eg they had a caseload of like 1-3 games) then I'm sure they could figure this out on a case by case basis. But I doubt they're willing to make that kind of investment.
But none of that makes it right, of course. I'm ambivalent about Valve's decisions here.
fnord already mentioned this, but to repeat: the situation with AI is the same as with other assets - Valve has no guarantee a developer didn't infringe copyright in taking some other model/texture/audio, tweaked it a bit, and put it in their game.
Edit for reply/clarification: "it's more likely" - I'm not speculating on the internal politics at Valve. Just stating the fact that, regardless of if assets are AI-generated or not, Valve cannot be certain they don't infringe copyright. In this respect, AI-generated assets with assurance the model wasn't trained on copyrighted works, are no worse than traditional ones.
it's more likely that there's an anti-ai evangelist on the team doing these reviews, and they are responsible for causing these issues (and using the legal turmoil surrounding AI copyright to hide their bias).
they know that because steam is such a large store, they can influence the direction of the game dev community and set a norm that they prefer to see.
because the copyright infringement of art assets were never a big deal and would only surface when complaints are sent to steam about certain games that copied assets.
So why all of a sudden, that ai generated assets would get such attention?
Because it's well established how copyright works with human-authored assets and unsettled how copyright works with generative AI, and corporations are allergic to uncertainty and liability? Along with the risk that they'll get overwhelmed with low quality submissions like eg Clarkesworld?
Given that in the past that have been very reticent about taking down games (eg see [1][2]), I don't find it likely that they have an "anti AI" stance. I personally would need to see like, a blog post or some other statement to that effect in order to buy that hypothesis.
Let me make it more clear, i am not saying valve's corporate stance is an anti-ai one. I am making up a conspiracy theory that there's staffer at valve who is anti-ai on the review team, and they are using their position as a reviewer to push an anti-ai agenda under the guise of (potential) copyright infringement. My only evidence is that the pre-existing reviews of games don't put heavy emphasis on proof of copyright. Nothing should've changed whether the production of the game's assets came from ai generation, or human generation, because the reviewer should be neutral and unbiased.
Well. Okay. I personally don't believe in conspiracies when there's an alternative explanation that doesn't really seem to have significant problems with it (corporations are hypocritical all the time) and when there's no real evidence of a conspiracy. If there were like, leaked emails or something I would reconsider, but until that time I don't see why I would prefer that hypothesis.
> There's no reason to suspect a generic asset is infringing on it's face, but for models trained on the open internet it's an entirely unanswered question.
As in, there's no reason not to give the benefit of the doubt to a traditional asset, but it's not yet understood how copyright works with a generative asset.
Caves of Qud already has in-game books generated by the old Markov chain technique, with the model trained on who-knows-what. It's on Steam, for now...
I think it's interesting to wonder what Valve would say about games using AI exclusively server side. They cannot control (or know) what technologies a developer uses on their personal game servers.
I think this game got unfairly caught in the crossfire (a sort of dead weight loss of consistently enforcing a policy), but it's easy to imagine Steam drowning in low-effort, slapdash games cranked out using AI. Like the junk you see on mobile, where the games are often "minimal viable ad platforms" rather than a "minimal playable games" (and one would hope for more than the minimum, anyway).
Though the top level comment asserting this is probably more about copyright than quality is probably right as well.
Steam is already drowning in low effort trash games. They don't surface much, since being a paid store, the ranking based on sales is pretty accurate to the quality of the game (not to mention the generous 2h refund).
Steam is already hard to browse because of a sea of low quality shovelware. AI, if that's what the comment is referring to, is likely to make that situation much worse and degrade the platform. And Steam gets that.
Steam started their current "we allow absolutely everything" policy because they were tired of people making fun of them for constantly banning Japanese games because they think anything anime-style is child porn.
Except the actual policy implementation is that Japanese games still get banned, even if they're all-ages and on every other storefront, but now it's also full of Western hentai games.
(This is consistent with Steam getting employees from Microsoft, DirectX/Xbox were codenamed on the theme "what's the most racist thing we can say about our Japanese competitors".)
> Steam is already hard to browse because of a sea of low quality shovelware. AI, if that's what the comment is referring to, is likely to make that situation much worse and degrade the platform. And Steam gets that.
Wouldn't you say there is a different between "Games made by AI" and "Games that included AI in one build for a specific feature as a test, and removed it"? The shovelware you talked about is the first, not the second.
On desktop it seems to be an even worse version of Reddit than the default (or old.reddit.com). Three columns of 'content' and the main post text (arguably the most relevant part of the page) isn't even completely visible.
(missing reddit less and less each day...)
Should be no difference. It looks like different languages. There's de.reddit.com for German, in.reddit.com for Indonesian. SH is Saint Helena I guess, which is a British overseas territory, and thus English. Not sure but that's my guess.
Would be fun and cool if submitter was linking to the Saint Helena version. I wonder what the odds would be, HN submitter + one person of ~4000 people living there, out of almost 8 billion
Call me cynical but my immediate impression that the gamedev is lying and exploiting the perceived asymmetry between the indie dev and the behemoth that is Steam. Attaching your game to controversy around AI is a great way to generate discussion and serves as a form of guerrilla marketing.
Maybe if the rules are "your game shall not contain AI", then don't write a ChatGPT integration for it.
51 comments
[ 5.0 ms ] story [ 108 ms ] threadI explained exactly how this worked and that there was no AI-content directly in the build, and even since then issued a new build without this mod ability just to be super safe. [...] Then, today, I randomly received an email that my app has been retired with a generic 'your game contains AI' response.
https://old.reddit.com/r/gamedev/comments/167iied/the_game_i...
I’m worried Valve will strangle what I think is the most exciting use of AI - artist trained models generating NPC content, lore, and procedural game content to create rogue like RPG of arbitrary complexity.
My guess would be that they can't verify what AI you used, at least without investing the kind of resources required to actually make case by case decisions.
I'm not saying this is your position, this is somewhat tangential, but I am confused why I seem to hear from people that both AI is a force of nature and companies stand in it's way at their peril, but also that it's a delicate flower and that any kind of management may strangle it in the crib. To be clear I know you aren't making both claims, I don't want to put words in your mouth, but it does feel like it's coming from the same general community of people. And I'm perplexed by that.
But yeah, I would also be interested to play and perhaps developing games like that. If someone is developing a game like that and they wants to hire a backend dev with experience integrating with models in a realtime context, shoot me an email. (I'm half kidding, I don't expect someone like that to happen to be in this particular thread, but only half.)
How do they verify copyright material isn’t in static assets? I can’t imagine it’s very thorough. I imagine most of it is based on attestation.
Regarding your question, I don’t share those views but can understand how they exist simultaneously. With an open hand, like was afforded internet and technologies, it could shape the world in hard to imagine ways as did the internet. However like the early internet it could have been strangled in the crib through knee jerk regulation and prohibitions. I think AI model builders and hosters want safe harbor similar to content hosters and providers allowing the end use of the technology to be regulated but the providers to be allowed to operate on a best effort basis. For instance the copyright question here is strangling any game that uses a generative model. A reasonable legal outcome is that it’s fair use to incorporate copyright material, but the output must not violate copyrights.
I’d note they also have no way to reasonably unpack all game assets, visual, text, and audio, and verify there’s no copyright material. To my memory there are several cases of video game publishers including copyright violating assets. It seems like a copout to take on faith a no copyright violation attestation for static content but not on models.
I'm sure you're right that's all they would need to win a lawsuit, but they would probably prefer to avoid one altogether. Or at least not to be the first publisher sued.
There's no reason to suspect a generic asset is infringing on it's face, but for models trained on the open internet it's an entirely unanswered question.
Indiscriminate crawler trained models are one thing, but models trained by the game developer directly should be handled differently. In that case they actually are the same as a generic asset, as they were assembled and curated by the game developer with an understanding of copyright compliance. But there are cases where they didn’t - particularly with audio.
If you tasked a human being with sufficient knowledge, latitude, and resources to investigate a game with a significant fraction of their time (eg they had a caseload of like 1-3 games) then I'm sure they could figure this out on a case by case basis. But I doubt they're willing to make that kind of investment.
But none of that makes it right, of course. I'm ambivalent about Valve's decisions here.
Edit for reply/clarification: "it's more likely" - I'm not speculating on the internal politics at Valve. Just stating the fact that, regardless of if assets are AI-generated or not, Valve cannot be certain they don't infringe copyright. In this respect, AI-generated assets with assurance the model wasn't trained on copyrighted works, are no worse than traditional ones.
they know that because steam is such a large store, they can influence the direction of the game dev community and set a norm that they prefer to see.
So why all of a sudden, that ai generated assets would get such attention?
Given that in the past that have been very reticent about taking down games (eg see [1][2]), I don't find it likely that they have an "anti AI" stance. I personally would need to see like, a blog post or some other statement to that effect in order to buy that hypothesis.
[1] https://www.thegamer.com/valve-gabe-newells-libertarian-stea...
This is summarizing reporting from People Make Games:
[2] https://www.youtube.com/watch?v=s9aCwCKgkLo [48m]
> There's no reason to suspect a generic asset is infringing on it's face, but for models trained on the open internet it's an entirely unanswered question.
As in, there's no reason not to give the benefit of the doubt to a traditional asset, but it's not yet understood how copyright works with a generative asset.
Am I missing something?
Except if you have the rights to all the works used to train your AI model. In that case, there is no risk that you are infringing.
I think it's interesting to wonder what Valve would say about games using AI exclusively server side. They cannot control (or know) what technologies a developer uses on their personal game servers.
[1] https://news.ycombinator.com/item?id=37370934
Though the top level comment asserting this is probably more about copyright than quality is probably right as well.
I was attacked for saying that entry barrier of 100$ was too low, but clearly that would have been only reasonable way to remove some.
Steam started their current "we allow absolutely everything" policy because they were tired of people making fun of them for constantly banning Japanese games because they think anything anime-style is child porn.
Except the actual policy implementation is that Japanese games still get banned, even if they're all-ages and on every other storefront, but now it's also full of Western hentai games.
(This is consistent with Steam getting employees from Microsoft, DirectX/Xbox were codenamed on the theme "what's the most racist thing we can say about our Japanese competitors".)
Wouldn't you say there is a different between "Games made by AI" and "Games that included AI in one build for a specific feature as a test, and removed it"? The shovelware you talked about is the first, not the second.
And no, it's not related to language.
Maybe if the rules are "your game shall not contain AI", then don't write a ChatGPT integration for it.
AI is such a broad term though. Either ban GPT/LLM architecture if that's the goal, or image synthesis, but be specific!
Probably more than 50% of all games out there contain something one could call "AI".
Then remove all the games I guess? :)
Where can I go to get games that do integrate AI/AI content?
We have good decentralized protocols these days.