> ## Content Index
> Fetch the complete content index at: https://www.regenraygun.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# Itch.io activates but Valve is still thinking about AI
- URL: https://www.regenraygun.com/itch-io-ano-ai-filter/
- Published: 2026-11-05T17:49:37.000Z
- Updated: 2026-11-05T17:50:06.000Z
- Description: Other platforms are taking a stand. Valve is (un)gatekeeping.
- Author: Jon Jordan
- Tags: Steam

A developer was told: pay up, or we will clone your game with AI and release it first.

That threat is what brought Valve to the *sujet du jour* at Steam Indie Connect, where Tom Giardino and Marc Wilkinson [addressed the cloning problem](https://www.gamedeveloper.com/pc/valve-addresses-steam-ai-cloning-problem?ref=regenraygun.com) for the first time. Valve's advice to the extorted developer was to report it to Steam and contact law enforcement. Its advice to everyone else was to build a community and keep updating the game.

On the broader question, however, Wilkinson said the company is fully aware of the situation and that a significant number of people are thinking about how to solve this problem. Giardino said Valve [doesn't want to be gatekeepers](https://www.pcgamer.com/software/ai/valve-admits-steam-has-an-ai-slop-and-cloning-problem-but-says-we-dont-want-to-be-gatekeepers-telling-developers-which-tools-they-should-or-shouldnt-use/?ref=regenraygun.com) telling developers which tools they should or shouldn't use. No new rules were announced. Wilkinson conceded the trade-off directly: some low-quality games get exposed.

So the largest storefront in games has acknowledged the problem and declined to do anything about it. Other companies are taking a different approach, however. 

On 3rd November, Itch.io [announced it had deleted a couple of terabytes of AI slop](https://www.pcgamer.com/software/ai/just-deleted-a-couple-terabytes-of-ai-slop-itch-io-pushes-back-against-the-flood-of-ai-on-its-storefront/?ref=regenraygun.com), extended its disclosure requirement from game assets to tabletop, books, comics and soundtracks, and [added a No AI filter](https://www.engadget.com/2305699/itchio-adds-a-filter-for-games-with-gen-ai-disclosures/?ref=regenraygun.com) to browse and search. 

Applying it took the catalog from more than 1,770,000 projects to fewer than 425,000.

The numbers are the clearest statement of the problem anyone has announced. Itch.io says the number of game pages created daily has doubled over the past two months, while downloads and purchases have not. Asked why, it replied: take a guess.

But a filter built on self-reporting inherits the honesty of the people filling in the form. Research puts actual studio AI usage at 50% versus 20% disclosing it. Itch.io has built a good tool on a foundation it cannot verify.

In contrast, Pantaloon certifies a judgement, and charges for it. [The Playbill launched on 22rd October](https://www.engadget.com/2294854/a-new-subscription-service-promises-an-evolving-collection-of-bold-and-bizarre-indie-games/?ref=regenraygun.com) with thirteen curated indie games, at five dollars a month, and a flat commitment against generative AI in any form. Nobody is self-reporting to Pantaloon CEO Jamin Smith. He vets, and puts his name on it. 

Realistically, the subscription buys the vetting rather than the games, which makes it the only model here where trust is the product rather than a cost of doing business.

Anthropic certifies nothing, but that's because it's intervening before anything exists. Around 27th October, it pushed a rule into Claude Code instructing the model not to recreate a specific existing game, even if renamed. 

[Addy Osmani confirmed it](https://x.com/addyosmani/status/2116213240363729013?ref=regenraygun.com): two sentences, added because using Claude to infringe IP breaches the usage policy. No disclosure mechanism is required when the output never gets made. Of course, it fails in its own way instead. [Tom's Hardware has confirmations](https://www.tomshardware.com/tech-industry/artificial-intelligence/anthropic-confirms-new-claude-code-limits-to-prevent-game-cloning-users-report-opus-5-5-is-also-actively-blocking-reverse-engineering-tasks?ref=regenraygun.com) of Opus 5.5 refusing unrelated reverse-engineering work, a studio reports being blocked on board-wargame modules commissioned by paying customers, and anyone cloning a game commercially can simply use any other AI model than Opus 5.5\. 

Which brings it back to Valve.

Steam has the disclosure data. Its form was [revised in January](https://www.pcgamer.com/software/ai/steam-updates-ai-disclosure-form-to-specify-that-its-focused-on-ai-generated-content-that-is-consumed-by-players-not-efficiency-tools-used-behind-the-scenes/?ref=regenraygun.com) to cover player-facing AI content while explicitly exempting internal workflow tools, including AI coding assistants — a sensible distinction that itch.io's broader net does not draw. Steam does not reject, delist or rank down a game for disclosing. It just collects the information.

What Valve has promised however is [smarter discoverability of high-quality games](https://frvr.com/blog/news/valve-wont-filter-out-ai-games-steam-but-a-significant-number-of-people-are-working-on-smarter-discoverability-of-high-quality-games/?ref=regenraygun.com), which is the one problem nobody has solved in this new era of abundance. 

And there is a case for indecision. Declarations are unverifiable, curation does not scale, supply-side vetoes catch the wrong work and miss the right people, and certification is meaningless where everything qualifies. A platform of Steam's size picking inaccurately would do more damage than one that waited.

But waiting has a cost too, and it is landing on the people Valve told to build a community. [Developers are delaying their games](https://kotaku.com/devs-are-being-forced-to-delay-their-games-on-steam-and-they-think-its-because-of-the-ai-slop-flood-2000752839?ref=regenraygun.com). Studios have cancelled public playtests rather than show work that might be cloned before release. Those are the costs of an unsolved problem being absorbed by the smallest parties in the market while the largest one thinks.

Other platforms have decided what they are prepared to vouch for. Valve has decided only what it is not: a gatekeeper. That is a coherent position, and it is also the one that leaves everybody else holding the risk.