Anthropic has updated the instructions for Claude Code, its AI programming tool, to discourage users from creating copies of existing games. The change follows growing concern among developers about AI being used to reproduce games without authorization.
GamesIndustry.biz reported that the updated guidance was identified on October 27 and applies to the Opus 5.5 model. It permits original games, projects based on established genres, and titles built from users’ own ideas and materials. However, Claude Code is instructed not to recreate a specific game, even if a request gives it a different name or describes the result as an original creation.
Anthropic technical team member Addy Osmani confirmed the change in a post on X, saying the company added two sentences to Claude Code’s instructions to guard against reproducing games protected by intellectual property rights, which would conflict with its usage policies.
The change comes amid broader industry concerns. Digital Sun recently postponed a public test of its game ReVamp over fears that the demo could be used as a reference for AI-generated copies. There are also questions about how platforms such as Steam might respond to a potential increase in AI-made games that reproduce other developers’ work.
It remains uncertain how effectively the new guidance can prevent game clones. Game Developer notes that safety instructions may be bypassed with certain prompts, and the tool may not be able to verify whether users have the rights to materials they provide.
The update is not a blanket ban on making games with AI. Original projects and games drawing on familiar conventions, such as platformers, role-playing games, and strategy titles, remain permitted. Anthropic says it plans to keep refining its rules to support creative use while protecting other developers’ intellectual property.
Sources: Olhar Digital.