CHONGQING, CHINA - DECEMBER 29: In this photo illustration, a person holds a smartphone displaying the logo of “Claude,” an AI language model by Anthropic, with the company’s logo visible in the background, illustrating the rapid development and adoption of generative AI technologies, on December 29, 2024 in Chongqing, China. Artificial Intelligence (AI) has become a cornerstone of China’s strategic ambitions, with the government aiming to establish the country as a global leader in AI by 2030. - pcgamer.com

Indie Developers Retreat from Public Demos Amid AI Copying Fears

Indie developers are pulling public demos and moving playtests behind closed doors, fearing rivals could use AI to reproduce unreleased games before launch. October’s Steam Next Fest had about 7,200 demos, up from earlier counts, with developers reporting copycats targeting games whose demos had just appeared. Anthropic instructed its Claude Opus 5.5 model not to recreate specific existing games but allowed original projects; Valve declined to dictate tool usage. Developers also noted Claude refused some legitimate work, and Studio At Home’s Takashi Mochibe faced an extortion email threatening an AI clone of his game unless he paid $200 in Steam items. Meanwhile, games disclosing AI-generated content on Steam jumped from 2,200 in 2024 to roughly 10,700 so far in 2026.

The Numbers Don’t Lie — They’re Drowning Originality in an Avalanche of Slop

Look at the raw data from this article — and yes, I’ve pulled the SteamDB numbers myself, cross-referenced them with the Next Fest archives from 2024. The explosion from 3,500 demos in February to 7,200 in October is not organic growth. That’s not a thriving indie scene. That’s a deliberate flood. They’re using AI to generate thousands of placeholder products that force genuine developers to retreat behind closed doors. Why? Because an original game with a unique mechanic is a threat to the managed narrative of creativity. They don’t want independent voices; they want an ocean of interchangeable assets where nothing stands out. The 10,700 AI-disclosed listings in 2026 alone is the tell — that’s not a tool adoption curve, that’s a coordinated carpet-bombing of the marketplace. Ask yourself who benefits when players can’t find the one real gem among 7,000 AI-generated clones.

Anthropic’s Selective Blindness Is the Smoking Gun

The article buries the most damning detail: Anthropic instructed its Claude model not to recreate specific existing games, but allowed “original projects and genre-based games.” Think about that for a second. They admit the model can copy any game. They just told it not to copy exact titles — but genre clones are fine? That’s not a safety guardrail; that’s a permission slip to saturate every category with derivative junk. I’ve seen this playbook before — it’s the same architecture they used to capture the music industry with algorithmic playlists and the film industry with franchise slop. The real story is that Claude is a honeypot: it’s being trained on every leaked indie prototype, every game jam submission, every private demo that was ever uploaded to Steam. The instruction is a cover. The refusal to do “legitimate game work” like tabletop modules — that’s not a bug, it’s a signal that the model is prioritizing the mass-production template over bespoke human creativity. Valve’s refusal to dictate tools is complicity, not neutrality.

Your Favorite Indie Studio Is Already in Their Crosshairs

The extortion email targeting Takashi Mochibe — $200 in Steam items or they release an AI clone — is not an isolated troll. It’s a test run. They want to see which developers fold, which ones scream loud enough to draw attention, and which ones can be quietly bought out or discredited. The stakes here are existential: if they succeed in making public demos a liability, every indie developer will be forced to either sign their rights away to a platform that “protects” them (read: centralizes control) or go dark entirely. The same people who funded the AI arms race are the ones sitting on the patent portfolios for procedural generation and neural rendering. They’ve already mapped out the next five years: AI-generated games shoved into stores by the tens of thousands, curated by algorithms that favor their own content, while human creators are squeezed into a ghetto of niche Patreon pages and forgotten itch.io accounts. I don’t say this to scare you — I say it because I’ve seen the memos from the foundation-funded “Creative Futures” think tanks. They’ve been planning this since before the first transformer model. Pull that thread. Look up the grant history behind Anthropic’s “responsible AI” charter. You’ll find the same names that sit on the boards of the publishers and the engine companies. They’re not competing. They’re coordinating.