Hyderabad: Artificial Intelligence has made many complex and time-consuming tasks easier across the world. However, it has also created serious new risks. One of the latest concerns is affecting small video game developers, who are now living under a strange fear.
Several independent developers have received emails from unknown people threatening them: “Pay us money, or we will create an AI-generated clone of your game and launch it in the market.” Frightened by these threats, several studios have removed their free demos from Steam.
Why Are Developers Removing Demos?
According to reports, such emails have been arriving for the past few months. In late September, Weasel Pit, the developer of Iron Pilgrim, claimed that they were asked to send $200 worth of Steam items within 24 hours. However, some reports have also alleged that the same developer had earlier issued a similar cloning threat to another developer, making the case somewhat controversial.
South Korean studio YNOT received a similar threat for their game Spookitchen. Following the email, the studio removed its demo on October 20.
Game developers usually upload demos on Steam so that players can try the game. Unfortunately, the same demos also allow potential cloners to understand the game’s look, feel and mechanics — information that is then used for blackmail.
What Anthropic Actually Changed
Around October 27, a user on X pointed out a new rule that appeared in Claude Code. On October 30, Anthropic’s technical staff member Addy Osmani confirmed that two new sentences had been added to the prompt of Opus 4.5 (referred to in some reports as 5.5).
The purpose of the change is to stop the AI model from recreating someone else’s specific game, as doing so would violate Anthropic’s usage policy. According to the company, work on original games, games in the same genre, and games based on the user’s own design should continue without restriction.
Several publications reported that Claude had completely banned reverse engineering. Anthropic has clarified that this is not the case.
Addy Osmani stated:
“We added two sentences to the Claude Code prompt on Opus to prevent it from recreating someone else’s specific game, because using Claude to violate others’ rights goes against our rules. Original games, games in a genre, and games based on the user’s own design should continue to work. We will keep refining to strike the right balance.”
The statement does not say that Claude has banned reverse engineering entirely. It only prevents the model from cloning a particular existing game. Some users have reported difficulties with crash dump analysis or reverse-engineering tools, but these are individual claims. Anthropic has said it will continue improving the balance.
How Real Are the Threats?
While developers have received threatening emails, no finished AI-generated clone of any game has been publicly confirmed so far. A Korean report noted that one of the emails contained neither a video nor any attachment. The same report also claimed that the developer who presented himself as a victim in the Iron Pilgrim case had earlier threatened another developer in a similar manner.
Valve’s Tom Giardino said he was hearing about such emails for the first time. The company does not want to restrict the tools developers choose to use. His advice to developers is clear: do not rely on just one simple mechanic. Add your own style, unique features and original design so that the game becomes harder to copy.
As AI tools become more powerful, the line between legitimate inspiration and outright cloning is growing thinner. Small developers are currently the most vulnerable, and the recent wave of threats has forced many of them to take defensive steps such as removing demos from public platforms.
