Thursday, June 25th 2026
Epic Games CEO Calls Steam AI Disclosure "Irresponsible"
Epic Games CEO, Tim Sweeney, has a reputation for his hot takes on the gaming industry, and he has previously been vocal about both the perceived benefits of AI in gaming and his criticisms of its detractors. Now, the executive has expanded on his prior comments about Valve's mandatory generative AI disclosure, which he previously said "makes no sense for game stores," in a new interview with PC Gamer. These comments follow the recent announcement that Unreal Engine 5.8 and Unreal Engine 6 will add more generative AI features to the game engine. Addressing the question of AI art and how it may be more of a PR problem, Tim Sweeney said that "it's just a question of adaptation of tools, because there's nothing like a prompt-to-game solution on the horizon that anybody expects will work," going on to explain that artists at Epic Games are already using generative AI to "reduce the drudge work," referring to some of the more tedious parts of the 3D modelling and texturing process.
He goes on to advocate for using AI, stating that "it's impossible to reconcile the idea that developers shouldn't use productivity-improving tools with the state of the industry." He goes on to explain that how you go about creating an in-game asset isn't important, and that "you might spend a million dollars modeling a flower pot to get the most detailed flower pot creatable by humanity, but you could also scan a flower pot with a very high resolution camera," and that "the value is not in creating a perfect flower pot...the real value is in building the scene and building the game and building the narrative." Ultimately, his argument is that AI makes it cheaper to make games and in-game assets, but he argues that "it's unfortunate that so many developers are put into this position. If you want to launch a game and get it as widely publicized as possible, you've got to put 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." He calls Valve's AI disclosure "irresponsible" and says that it makes it "much, much, much harder for a game developer to have a chance of success."Sweeney's comments follow a recent analysis that saw AI games on Steam sell as many as 53% less units, so his argument that the tech deters players may hold water, however, it should also be mentioned that the aforementioned analysis also found that the backlash was worse when it was a large studio using AI, and that players frequently reviewed games made with AI more negatively. Anecdotally, there have also been several instances of backlash against successful games using AI art for final assets or during the development process, and several indie game developers have spoken out against the use of generative AI, including Pocketpair and Fireshine Games. The developer of Vampire Survivors also recently announced that it would be reconsidering its collaboration with Fortnite after Epic Games showed off an AI-heavy workflow on social media.
Despite his positive attitudes towards AI art, Tim Sweeney seemingly does not think that AI can replace artists in the game development pipeline, with the CEO commenting that "You're not going to create good art by giving a computer a prompt and having it spit out a mesh. You're going to get good art by an awesome artist using all of the tools available to make it."
Source:
Tim Sweeney via PC Gamer
He goes on to advocate for using AI, stating that "it's impossible to reconcile the idea that developers shouldn't use productivity-improving tools with the state of the industry." He goes on to explain that how you go about creating an in-game asset isn't important, and that "you might spend a million dollars modeling a flower pot to get the most detailed flower pot creatable by humanity, but you could also scan a flower pot with a very high resolution camera," and that "the value is not in creating a perfect flower pot...the real value is in building the scene and building the game and building the narrative." Ultimately, his argument is that AI makes it cheaper to make games and in-game assets, but he argues that "it's unfortunate that so many developers are put into this position. If you want to launch a game and get it as widely publicized as possible, you've got to put 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." He calls Valve's AI disclosure "irresponsible" and says that it makes it "much, much, much harder for a game developer to have a chance of success."Sweeney's comments follow a recent analysis that saw AI games on Steam sell as many as 53% less units, so his argument that the tech deters players may hold water, however, it should also be mentioned that the aforementioned analysis also found that the backlash was worse when it was a large studio using AI, and that players frequently reviewed games made with AI more negatively. Anecdotally, there have also been several instances of backlash against successful games using AI art for final assets or during the development process, and several indie game developers have spoken out against the use of generative AI, including Pocketpair and Fireshine Games. The developer of Vampire Survivors also recently announced that it would be reconsidering its collaboration with Fortnite after Epic Games showed off an AI-heavy workflow on social media.
Despite his positive attitudes towards AI art, Tim Sweeney seemingly does not think that AI can replace artists in the game development pipeline, with the CEO commenting that "You're not going to create good art by giving a computer a prompt and having it spit out a mesh. You're going to get good art by an awesome artist using all of the tools available to make it."

94 Comments on Epic Games CEO Calls Steam AI Disclosure "Irresponsible"
But as ive said, efficiency isn't an important metric for most people when buying pc hardware
You're overestimating how willing people are to pay for personal software outside of affluent areas in 1st world countries. The reality is much less glamorous than what you're painting.
As for myself, I've been a Linux user for many years and I just use Blender 3D and other FOSS applications, mainly for 3D character sculpting/modeling at a hobbyist level. But Blender hasn't always been good in the past (up to versions series 2.8-2.9) and still requires significantly higher GPU resources than, say, ZBrush. So, the typical Windows amateur will just use that, regardless of what he can actually afford.
...
Putting this aside, despite free & open-source software being actually decent as of 2026 I expect that the number of 3D artists (or digital artists in general) will decline in the coming years as generative AI models (even for 3D models) improve and make investing a good chunk of one's life into artwork an even worse deal than it previously was.
helpx.adobe.com/photoshop-elements/desktop/download-and-install-photoshop-elements/photoshop-elements-downloads.html
www.sidefx.com/products/houdini-apprentice/
www.autodesk.com/products/maya/free-trial
www.autodesk.com/products/3ds-max/free-trial
lightwave3d.com/products/trial-versions/
downloads.foundry.com/modo/retired :D
etc. etc. etc.
And most of them are not feature limited nor ever were. A minority have time restrictions but most just use file formats incompatible with the commercial versions - fine for learning. The packages which use heavy feature limits or require student status are typically not industry stables anyway. And once again; people who think like pirates and thieves typically don't explore these options so likely are generally unaware.Yeah, I'm not talking about a poor village in the Philippines or Afghanistan but the axiom still holds: Name me a successful studio in one of those places or any studio which uses places like that as a talent pool - potential or real.I learned ZBrush at one point 20+ years ago but haven't followed it intensely. Only a few years after the popularity spike it became fairly irreverent - a niche tool for a niche audience. Currently I'm told, "no one [besides hobbyists?] uses it for anything serious".That's been true for awhile tho. Easy-street industry first adopter opportunities were pretty much gone by, ummm... 2005 or something. Right now and for the past 3 or 4 years, we're seeing drastic changes in both the tools themselves and typical production workflows. Is that a good thing? For some yes, for others no. Is it a bad deal or poor career move? For some yes, for others, no. Still, for those able to express themselves well artistically and market that talent effectively, it's very lucrative and likely will continue to be for as long as just economies exist.
And the ancient axiom will of course remain the case: people willing to work, pay their way, and investigate all the legitimate opportunities, will generally be those that succeed and people who make excuses for poor decisions, bad behavior and take shortcuts (like theft or piracy) will fall by the wayside. With very few exceptions...
Have you perhaps considered the possibility that professional software is so expensive that quite many startups and small businesses can't afford it?
Don't get me wrong, I'm not advocating piracy here, only the general idea of things not being black or white in life, and moral ambiguity sometimes being a genuine thing.
I don't know where your "lived experience" comes from. Mine comes from an area where not just regular people, but even small businesses don't have money to pay for software licences. Some don't even have money to pay their taxes properly. It doesn't mean that they're morally corrupt. It only means that such software companies don't set their prices to such markets where I come from.
Sometimes, necessity is a factor. Sometimes, morals are a luxury. Think about it.
But I think we're getting into the weeds with potentially irrelevant specifics which appear to be typical on forum thread exchanges about AI, in game dev, etc. - with the usual mix of pro-AI pragmatism, anti-AI skepticism, and philosophical tangents. :D
Perhaps a bit of recontextualizing would be helpful? I admit being more skeptical and nuanced on AI agency/ethics than @Solid State Brain with points about "human artists train on "pirated" or copyrighted material too" framing AI training as analogous. Arguing that modern digital artists largely self-train on copyrighted/pirated IP, courses, and tools and therefor AI scraping/training isn't uniquely immoral is to me whataboutism in it's worst form. This is a common "both sides do it" defense of generative AI and I just happen to disagree with the extent as well as the relativism. I want to call out distinctions around human sovereignty vs. distributed AI systems and reject direct equivalencies between individual human learning and machine training. I believe it's worth pushing for moral/philosophical differentiation.
I believe this mirrors the broader debate as well: transparency vs. stigma, human creativity vs. tool augmentation, and whether labeling AI use is "irresponsible" fearmongering (Sweeney's view) or basic consumer info (Valve's policy) - and thus brought it to light.
Valve is right on disclosure. Consumers deserve to know if a game's assets (art, music, code, writing) lean heavily on generative AI, especially in an era of flooding marketplaces with low-effort slop - AI misuse and overuse. It's not "hate", it's informed choice, like nutritional labels or "made in X" tags. Sweeney's position smells like self-interest: Epic/Unreal is doubling down on AI integration (UE6), and disclosure could hurt sales for titles perceived as "AI-slop." Tim loves fighting Valve, but this feels like protecting the pipeline rather than artists or consumers.
The artist training analogy is directionally true but incomplete and ill-proportioned. Yes, humans "train" on vast copyrighted works (most often via fair use, education, paid, and even occasionally gray-area piracy). But scale, speed, and economic displacement differ massively. One human artist internalizes influences over years and produces original synthesis; a model ingests billions of works almost instantly to output derivatives on demand. The output volume and job market compression are the real disruptions - potentially. Equating them too cleanly dismisses legitimate concerns about consent, credit, and compensation for original creators whose data are powering the models.
I still think the sovereignty point is interesting. My angle on individuals being "sovereign" vs. AI as distributed/impersonal hits the crux of the unease most people seem to deliberate: AI feels like an aggregated cultural commons being repackaged without agency or reciprocity. Humans create under personal risk/reward; models don't. That said, tools evolve - i.e. photography didn't end painting, but it did change it dramatically.
"Irresponsible" is peak Sweeney hyperbole. Valve's policy is mild and factual. The real risk to devs is backlash from vocal anti-AI crowds (some of whom are hypocritical or just plain Luddites), but hiding AI use breeds distrust. It's way better to disclose, market the human elements strongly, and let quality win. Games, films and other forms of digital media with thoughtful AI augmentation (e.g., procedural help, not full asset replacement) can succeed if the final product is good. And of course I think we all truly believe pure AI shovelware should be called out.
I think this is healthy forum friction so far - much better than echo chambers. Pro-AI folks often underrate cultural/ethical externalities; critics often overrate "soul" while ignoring how all creative tools (brushes, DAWs, engines) automates what was priorly laborious. I think the winner will be hybrid approaches which amplify human vision and creativity rather than replace it wholesale. Epic pushing hard on AI while criticizing disclosure feels inconsistent (at best) if their goal is truly better games for everyone.
Again, I'm not advocating piracy. I'm only advocating a non-judgemental attitude and an open mind. The world isn't as black and white as you seem to think. And the people we're talking about certainly don't need your opinion.
Or, the "birds of a feather" adage applies and certain people have shared experiences very different from the main stream - that could of course apply to either side in this discussion. I'm am however, going to continue to believe that the majority of the industry I spent my life in isn't filled with selfish and corrupt corner-cutting thieves, and that part of the template for success is earning your way through hard work and dedication. I really don't know too many lazy thieves who have made it. Do you?
You're also dead wrong about several of your assumptions. Free versions as I said in my previous post have been a thing for multiple decades! And I don't know your age but you were likely in diapers (82?) when I began teaching the subject. Schools were both far less expensive and fully equipped with current hardware and software. I'm still on the advisory board for one university system here and they still are. Pirates were never welcome and still generally aren't. Even letting that be known generally excluded you from consideration at most if not all studios I've ever hear of. No one wants to risk working with people who can justify thievery because of either one of the lame excuses being presented here: "it's to hard to spend a month or two's salary on the tools needed" or "I'm just too lazy to look for educational/free versions and they might not have 100% of the tools anyway".
Whether you believe it or not, people not having money "to do the right thing" is a real phenomenon. The wallet of a rich software company's CEO is of lesser concern when you can barely bring your small business to net zero even without paying for unaffordable software licences.
You want to be a sound engineer here? What do you think is needed in order to learn on your own? I only know my own experiences and most of that is hobby level so I aksed AI. The answer made good sense to me so here's that:
[INDENT=2]"Here's a solid, no-frills beginner-to-intermediate sound engineering setup focused on recording, mixing, and critical listening. This is practical, reliable, and widely used by pros for home/studio work. Prices are current approximate US retail (as of mid-2026; check retailers like Sweetwater, Amazon, or Thomann for deals/taxes/shipping). [/INDENT]
[INDENT=2][/INDENT]
[INDENT=2]Core Hardware[/INDENT]
[INDENT]
- Computer: Apple MacBook Air 13-inch (M4, 16GB unified memory, 512GB SSD) — Excellent stability, low latency, and performance for DAWs like Logic Pro or Pro Tools.
- Audio Interface: Focusrite Scarlett 2i2 4th Gen — Best all-rounder for beginners with great preamps and included software.
- Shure SM57 (dynamic) — The industry workhorse for instruments, amps, and more.
- Audio-Technica AT2020 (condenser) — Solid for vocals and acoustic instruments.
- Studio Monitors: Pair of JBL 305P MKII (5" powered nearfield) — Accurate, affordable, and great for mixing.
- Headphones: Audio-Technica ATH-M50x — Closed-back, neutral reference for tracking and mixing.
[/INDENT]≈ $999–1,299 (base configs often available around $999–1,099 on sale).
≈ $190
≈ $109
≈ $119
≈ $338 (pair)
≈ $159
[INDENT=2]Accessories (Essentials)[/INDENT]
[INDENT]
- XLR cables (2–3), mic stand, pop filter, and basic isolation shield: ≈ $80–100
[/INDENT][INDENT=2]Total Estimated Cost: ≈ $2,000–2,400 (depending on exact Mac config and deals). This gets you up and running professionally without excess.[/INDENT]
[INDENT=2][/INDENT]
[INDENT=2]Software (Start Here)[/INDENT]
[INDENT]
- DAW: Logic Pro (one-time purchase, Mac-native and powerful) or Pro Tools Artist/Intro (free tier to start, subscription for full features).
[/INDENT]Many interfaces include free plugin and bundled software.
[INDENT=2]This setup teaches proper signal flow, mic technique, and critical listening. Prioritize treating your room with acoustic foam/bass traps (extra ~$100–200) for best results. As you grow, you can expand I/O or upgrade mics/monitors. This is battle-tested and scalable—start recording and mixing immediately!"[/INDENT]
[INDENT][/INDENT]
Those are all new prices (current retail as of July 2026) and if purchasing used it's roughly about half that. Let's say it's a mix of used and new so $1,700. That is exactly the take-home pay you get working at the corner store for 4 weeks. So, now you have a choice: Couch-surf for four to six weeks and buy it all at once or keep your $300~$600 a month apartment (rent can be very cheap - in Japan just like the USA!!) and buy one or two of the items on that list every month. There is absolutely no need to steal anything - especially the software. Even if they decided to go with a Cadillac of DAWs like Steinberg's Cubase instead of one of the MANY free or cheap ones out there we're only adding at most $500 and more likely $60 for learning or even free with a hardware purchase - Steinberg has been offering free full featured "evaluation" versions (Cubase SX) since 2004 BTW - for those of you thinking this is something new - LOL.
OK, so now you have your gear - DON'T QUIT YOUR DAY JOB! Learn as much as you can on-line or in books - shouldn't take too long if you're aggressive (~2 months max?) and go beat the streets on your days off - still the best way to find a job even in 2026!
[INDENT=2][/INDENT]
[INDENT=2]"Current full version prices (USD, directly from Steinberg as of July 2026) for Cubase 15: steinberg.net[/INDENT]
[INDENT]
- Cubase Pro 15: $579.99 (full-featured flagship version)
- Cubase Artist 15: $329.00 (mid-tier, great for serious musicians/producers)
- Cubase Elements 15: $99.99 (entry-level, budget-friendly starter)
[/INDENT][INDENT=2]These are the standard retail prices from Steinberg's official store. Educational discounts are available (e.g., Pro ~$360, Artist ~$200, Elements ~$67 for eligible students/teachers). steinberg.net[/INDENT]
[INDENT=2][/INDENT]
[INDENT=2]Note: Prices can vary with promotions/sales (Steinberg and retailers often run discounts, such as 25-30% off). Upgrades from previous versions are cheaper than full price. Check the official site for your specific upgrade path and current offers. [/INDENT]
[INDENT]
- All versions share the same high-quality audio engine and core workflow.
- Cubase LE and AI are lightweight versions often included with Steinberg/Yamaha hardware. steinberg.net
[/INDENT][INDENT=2]You can download trials of the full versions from Steinberg's site to test them. For the most accurate/current pricing (including any sales or your upgrade eligibility), visit the official Steinberg Cubase pages directly."[/INDENT]
If your "friend" is struggling with those kinds of numbers there is something wrong - likely the result or consequence of some very poor choices - and this is no different than in the USA. Your excuses are getting more and more pithy as we go on here.
1. This person doesn't have a day job. He is a sound engineer, that's what he does for a living.
2. He can't even have a day job. With that, he wouldn't have money for any gear, or any time to do his engineering job. In my hometown, rent is more expensive than what a basic 8-16 job earns you. The only way to survive is to live with your family which being in his late 20s, he doesn't have. Not to mention mandatory overtime rules in the country which make it impossible to do anything else.
3. With such economic conditions, you don't buy your gear in 4 weeks. Like I said, he slowly got it item by item through years of bartering.
4. The minimum amount of self-employment tax is about $150 a month there. He doesn't have money to pay that regularly, that's why he's doing the job unofficially. What makes you think that he has money for professional software? What makes you think that saving up and paying for professional software is a priority when he can't even pay tax regularly? I'm sure he would be happy to pay for his software if he had a predictable, regular income... Which he doesn't.
If you can't be asked to learn about Eastern European economic conditions, then at least do the favour of keeping your prejudice to yourself. Thanks.