Steam Next Fest is upon us, and fans agree it's the AI-sloppiest one yet.
Next Fest, Steam's occasional multi-day promotion in which developers release demos of their games so players can get a preview of what's to come, has increasingly become unwieldy due to the sheer number of demos to play, but the advent of AI models and vibe-coding has exacerbated the problem. There are approximately 7,000 demos to wade through this time, almost doubling the total from the last Next Fest in July. It's an absurd, overwhelming number that makes the event, which used to be great for trying new, unexpected titles, feel practically meaningless.
As bad as the flood of demos is for a promotion that was once about discovering new games, though, the situation today is actually worse. Not only is Next Fest getting an increasing number of bad slop games, it's also costing us the opportunity to play some good, promising, handcrafted games.
When I saw the trailer for ReVamp, I was excited. It appears to be a clever and stylish tower defense game that puts you into the imposing leather boots of Dracula himself, as he builds his lair to withstand the onslaught of heroes aiming to defeat him. Its gameplay looks similar to Fallout Shelter's build mechanics, but with a new hook and the ability to personally intervene as Drac himself. You customize your very Castlevania-looking environment to serve as a labyrinth for heroes to pass through. In short, it looks rad as hell.
So I was disappointed this week to hear that ReVamp had canceled its plans for a public playtest, citing fears that a big public demo would be downloaded, scraped by AI models, and cloned.
"We've been concerned by recent news of games being reverse-engineered, cloned, and published before the original is even out," developer Digital Sun said in a statement. "With launch still months away, a large public build of ReVamp felt like too big a risk."
At the same time, Digital Sun acknowledged that this puts indie developers between a rock and a hard place--being unable to share their games for fear of them being cloned also means forgoing a major tool indie developers have historically used to garner word-of-mouth exposure before release.
"Showing games early is how indies get seen and find the fans who keep us going, so we hope platforms keep strengthening protections for original work," the studio said.
The statement did not specifically mention Steam's event, but it came alongside the week of Next Fest, so it's difficult not to make the connection. Digital Sun did say it will run a playtest through its Discord server later this month, which may offer more protections against clones with the game going out to a smaller community--but which also results in fewer people learning about the game.
It's a shame, but it's also hard to blame Digital Sun, or any other developer who chooses to pass on releasing a demo. The risk of getting beaten to the punch by a low-effort clone is just too great, and indie developers likely don't have the legal muscle to challenge what could be a cavalcade of lazy theft. Copies of popular upcoming games are nothing new, but the introduction of AI tools has vastly reduced the amount of time and work it takes to make one.
To show how easy cloning is and what a problem it presents, game developer Tim Ruswick cloned a game, with permission from the original developer, and showed the results in a YouTube video. The original game, Polyfury, was made in 2023, and Ruswick tried three different methods of cloning it, with varying results. Those ranged from providing Anthropic's Claude AI with a single sentence prompt, feeding it the Steam page, and even uploading the actual game code itself. In the comments, Ruswick mentioned that some of these were generated in as little as 30 minutes.
The results were mixed, and counterintuitively, the single-sentence prompt actually worked better in some ways than providing the AI with the full game code. None of the clones quite functioned perfectly or were as full-featured as Polyfury. Ruswick also mentioned that this experiment was meant to illustrate the ease and speed of producing clones, not producing them with 100% accuracy, and given a few hours or days it would be easy to make a perfect clone. Even at this speed, though, they're universally close enough to fool the casual onlooker, or to siphon interest from an original idea before it gets a chance to put its best foot forward.
So it's no wonder developers are increasingly hesitant to put their demos on Steam, especially months in advance of a release date. They just can't afford the risk that their momentum will get sapped by someone easily stealing their work. But the net effect is that events like Next Fest get that much worse. Not only are there a glut of AI slop games, but the advent of AI also risks chasing out quality crafted games from the event.
It's hard to see a solution for this outside of major systemic change. Digital Sun closed its statement with a wish for greater protections from platform holders. Certainly Steam could do more to separate the wheat from the chaff. But as long as the AI models themselves are allowed to run unimpeded, there's only so much Steam or any other platform holder can realistically do to stop bad actors from using AI to clone games.
Until there's some larger change to Steam or the AI ecosystem, we're in what feels like a doom loop for big demo events like Steam Next Fest: more incentive to publish low-effort crap, less incentive to publish quality handcrafted games. It makes the event less useful for both developers and fans. This could not only destroy the utility of events like Next Fest: It could be the canary in the coal mine for indie games more broadly.
If studios can't even risk participating in an event meant to boost discoverability, what does that mean for the viability of spending years developing a game, building word of mouth, and releasing it, only to have a knockoff version vibe-coded within days? Like many issues regarding AI adoption, it seems inevitable that large swaths of the industry will be harmed, until and unless there are major reforms--the sooner the better.