Anthropic has made significant adjustments to the usage policy of its Claude Code tool, which runs on the Opus 5.5 model. The new guidelines introduce a direct ban on recreating existing video games—even in cases where the user changes the project's name or attempts to present it as an original development. This decision marks an important stage in shaping the norms for the use of generative AI in the gaming industry.
The company's official position was voiced by developer representative Addy Osmani. According to him, the key motive was the protection of intellectual property amid the rapid growth in the number of AI-created copies of popular gaming products. The industry has faced a wave of projects that effectively duplicated the mechanics, visual styles, and plot elements of well-known titles, which created serious legal and reputational risks for developers of AI tools.
What remains permitted
It is important to emphasize that the restrictions are targeted and do not block legitimate creative work. Claude is still able to assist in the development of original projects, the creation of games within existing genres, as well as products based on the user's own materials. Thus, Anthropic seeks to maintain a balance between protecting the rights of rights holders and creative freedom for bona fide developers.
The company also stated its intention to consistently refine the restriction mechanism so that it does not create obstacles to legal development. This points to an iterative approach: the policy will be clarified as practical experience and community feedback accumulate.
Cryptalist analysis: Anthropic's move reflects an emerging industry trend—leading developers of AI models are shifting from reactive moderation to preventive intellectual property protection policies in content. For the market, this is a signal of the gradual institutionalization of rules for the use of generative AI, which reduces legal risks for corporate clients but at the same time increases regulatory pressure on the industry itself.