Hate Slop, Not AI

I get why people are tired of AI slop. Nobody wants to open a game store and spend their time sorting through nearly identical uploads, misleading screenshots, and projects that don’t deliver what their pages promise. Creators who put care into their work shouldn’t have to compete for attention with someone flooding a storefront with disposable content.

But I think we’re making a mistake when we treat any use of AI as evidence that a creator belongs in that second group.

itch.io recently reported that the number of new game pages created each day had doubled in two months, while downloads and purchases hadn’t doubled alongside them. AI may well be contributing to that increase. The figures alone don’t tell us how much of it comes from AI or how many of those pages represent spam. They do give us a reason to care about discovery. More pages are competing for attention, but downloads and purchases haven’t kept up.

Some of the proposed responses concern me. Kenney has suggested daily upload limits, further restrictions when all of those uploads disclose AI use, a no-AI filter enabled by default, and community notes on project pages.

I do see a case for limiting how quickly an account can publish new projects. Community feedback could also help people assess what they’re about to download. But adding a penalty because someone disclosed AI use makes the choice of tool a reason for punishment before we’ve established anything about the work itself.

What does “made with AI” actually tell us about a game?

A developer might write the story, draw the art, compose the music, and use a model to help implement an inventory system. Another might use generated textures in an environment they’ve spent months designing. Someone else might generate the dialogue, artwork, and voices, then publish the result with very little review. Those are very different creative processes, and a single label doesn’t tell us much about the work or the decisions behind it.

To be fair, platforms can define their terms, and some already provide more detail. itch.io’s disclosure system distinguishes graphics, sound, text and dialogue, and code. But it also assigns a common AI tag. A default filter that excludes the entire category would collapse those distinctions again.

Dexerto recently reported that an independent analysis found AI disclosures on nearly half of Steam’s October 2026 releases. Those disclosures covered uses including artwork, translations, music, code, and marketing materials. That tells us something about how people are making games, but it doesn’t tell us how much creative judgment went into them or whether they’re worth playing.

That’s where I think mandatory labels can do real damage. A creator could be honest about using AI for a small part of a project and lose visibility before anyone considers what they’ve made. Someone uploading disposable content could deliberately lie and avoid the same filter. Nobody is owed an audience, but a default exclusion is still a significant decision for a platform to make about whose work gets seen.

It also creates an incentive to say less about your process. If explaining how you used a tool means being hidden or treated with suspicion, openness carries a cost. I think everyone would much rather have an environment where creators feel comfortable showing their work, discussing their decisions, and explaining what they learned.

VRChat’s recent statement on AI illustrates the problem very well. It reports that 95% of its survey respondents wanted some kind of AI label. But across eight example categories, no single combination of what should be labeled attracted more than 10% of respondents. There was broad support for a label and much less agreement about what it should mean.

VRChat decided against AI labels and filters, citing dishonest labeling and the risk of harassment, including accusations against people who haven’t used AI at all. It has prohibited misleading thumbnails and plans to add world ratings and larger image galleries. I agree with that direction. Give people a clearer view of what they’re entering, improve how they find worthwhile experiences, and address behavior that misleads them.

I think it’s reasonable to want to know how something was made. Some people value a particular process as much as the finished work. They might want to support artists who work entirely by hand, or avoid generative AI because of concerns about training data, consent, or employment. Those preferences deserve respect, and a polished result doesn’t make every concern about its production disappear.

Specific explanations can help people make those choices. Knowing which assets were generated, what a creator changed, or whether a voice was used with a performer’s consent tells me far more than a general AI badge. I want useful transparency, and I don’t think a blanket label that limits a project’s reach is a good substitute for it.

We also need to be careful with the word slop. I don’t want it to become an excuse to remove every game that looks amateur, every strange experiment, or every first attempt that isn’t particularly good. An open platform should have room for people to learn. A simple game can still give people something worth playing.

Needle In A Haystack Simulator is a good example. Its simple premise went viral, and a wave of lookalikes appeared before the original had even launched. Its creator eventually joked about a “Find the Original Game” mode, where players would search through a pile of imitation game cases. What bothers me is people rushing out imitations to cash in on the attention someone else’s work attracted. AI can make that easier, but I’d object to that behavior however the copies were made.

Platforms need rules more concrete than whether someone thinks a project is slop. Nearly identical listings, deceptive screenshots, misleading descriptions, and accounts flooding a storefront with uploads give moderators something specific to investigate. Apply those rules consistently, whatever tools were used. Improve previews, search, recommendations, and curation so people have a better chance of finding work they enjoy. Community feedback can help, as long as there are safeguards against coordinated attempts to punish a creator for the tools they use.

AI can make producing large amounts of disposable content easier, and that creates a real moderation problem. I’m not suggesting platforms ignore it. I want them to address the volume, repetition, and deception directly, because excluding everything associated with AI also excludes people making thoughtful work.

As I wrote in my last note, nobody owes AI a place in their creative process. That freedom should extend to creators who choose to explore it. Using AI to help build a game, realize an environment, or attempt something beyond the resources of a small team doesn’t automatically erase the person’s creative contribution. There can still be substantial work in directing, selecting, revising, testing, and bringing those pieces together.

Creators remain responsible for what they release. Using AI doesn’t excuse misleading an audience, disregarding consent, or publishing work they haven’t bothered to evaluate, but it also shouldn’t make their work unworthy of consideration before anyone has experienced it.

I want fewer spam uploads and more room for people making things they care about. That includes people who use AI and people who want nothing to do with it. We should be able to protect that room without turning the tools a creator uses into a reason to demonize them.