Anthropic's Claude AI blocks video game cloning and reverse-engineering

Anthropic's Claude AI is reportedly blocking attempts at game reverse-engineering and cloning, prompting concerns among developers. Twitter user Chasm claims that the model is refusing to create original games, games in a genre, or games from the user's own designs and assets. However, there is no official confirmation of this rule text.

Reports of Claude Code blocking standard game work and cybersecurity-related tasks have been flooding its GitHub. Anthropic introduced a Cyber Verification Program on October 6, limiting access to its advanced cybersecurity capabilities to entities covered by Project Glasswing. The timing of these restrictions is puzzling, as reports of bot refusals only began appearing within the last week, despite the Cyber Verification Program being in effect for 24 days.

Experts suggest that Anthropic, along with other AI companies, may be taking preemptive measures to avoid potential lawsuits from software companies for enabling reverse-engineering of their software. Users are reportedly shifting to Chinese, open-weight, or local LLMs to bypass Claude or GPT restrictions. This situation highlights the irony of frontier AI companies copying the world's information and art for training sets while now fearing copyright infringement.

The recent controversy surrounding the Steam Next Fest and its AI-generated content has added fuel to the fire, with small studios facing extortion threats from clone makers.