1. News
  2. AI models & agents
  3. Games · AI scams
Games · AI scams3 min read

Small game makers get threats over AI copies of their games

Scammers are emailing indie studios, demanding money or saying an AI copy of their game will launch first. No working clone has been shown so far.

By the Dashline deskSource: ETV Bharat
Small game makers get threats over AI copies of their games
Photo: Alphaaloy, CC BY-SA 4.0

Scammers are emailing small game studios with a blunt threat: pay up, or an AI-made copy of your unreleased game will come out first. Demands run from hundreds to thousands of dollars. Out of fear, several developers have pulled free demos from Steam, the big PC game store. So far, no one has shown a fully working AI clone.

What the threats look like

Reports say the emails have circulated for several months. In late September, Weasel Pit, maker of Iron Pilgrim, said it was asked for $200 in Steam items within 24 hours. Some reports say the same developer sent a similar clone threat to another developer weeks earlier. That has led to a dispute over whether the claim is legitimate.

South Korean studio YNOT got a similar email about its game Spookitchen and pulled its demo on October 20. Demos give copycats a clear look at how a game looks, feels and plays. That gives blackmailers something to point to.

Little proof behind the emails

An Iven Global report from Steam Indie Connect says some of the emails had no video evidence or code attachments at all. Valve's Tom Giardino said it was the first time the company had heard of such threats. Valve calls it a matter for law enforcement, according to Risky Bulletin.

Anthropic changes Claude's rules

Anthropic, the company behind the AI assistant Claude, was pulled into the story. It added two sentences to the instructions for Claude Code, its coding tool, on the Opus 5.5 model. Users on X noticed the rule around October 27. On October 30, Anthropic staff member Eddie Osmani confirmed it. The aim is to stop Claude from rebuilding a specific existing game, which Anthropic says breaks its usage policy.

Original games, games in a genre and games from a user's own designs should still work. Some outlets reported that Claude had banned reverse engineering, which means studying software to see how it works. Anthropic denied that. Some users said they had trouble with certain reverse engineering tools, but those are their own claims.

Anthropic later said the rule also kicked in when people only named a game to describe the mechanics they wanted. It updated the rule in version 2.1.317. Now a user can name a game, and Claude should build an original game in that genre. If asked for a copy, it offers to build something similar.

What creators can do

  • Ask for proof, such as video or code, before paying anything.
  • Report the email. Valve treats these threats as a law enforcement matter.
  • Do not rest a game on one easy-to-copy mechanic, Giardino advised.
  • Build a distinct art style and deeper features that AI cannot easily copy.

What is still unknown

The emails are real, but no finished AI clone has surfaced. Anthropic's statement does not mention the extortion emails, so the two stories are linked only in coverage.

The takeaway

If you make games or anything creative, treat an AI copy threat with doubt, ask for proof before paying, and report it.

More articles

You should really read these.

Next step

Let’s talk