# Anthropic Restricts Claude Code from Recreating Existing Video Games New system prompt updates prevent intellectual property infringement while allowing original game development. - **Source:** [The Verge](https://www.theverge.com) - **Published:** 2026-11-02 - **Topic:** [Data & AI](https://brifing.si/c/data-ai) - **Original article:** - **On Brifing:** https://brifing.si/s/36a0141b-0221-42e1-9007-2b4b34dcbc1c ## System Prompt Update Strategy Anthropic has officially modified the system prompt for Claude Code on the Opus 5.5 model. This update explicitly instructs the AI to avoid recreating specific existing video games. The change addresses growing concerns regarding the use of AI tools to generate unauthorized decompilations, recompilations, and ports of established gaming titles. ## Defining Permissible AI Development The updated instructions distinguish between prohibited cloning and allowed creative work. Users are still encouraged to build original games, explore specific genres, or develop projects based on their own unique designs and assets. This policy shift aims to foster innovation while maintaining strict adherence to established intellectual property usage guidelines. ## Addressing Intellectual Property Risks Addy Osmani confirmed the update, noting that using Claude to infringe on third-party IP violates the company's usage policy. By embedding these constraints directly into the system prompt, Anthropic seeks to automate compliance. This proactive measure prevents the tool from being leveraged for direct clones of copyrighted software and games. ## Verification of Prompt Changes The change was identified after a user on X shared a screenshot of the new system prompt language. The added text explicitly forbids building recreations of specific existing games, even if the user labels them as original. Anthropic has not yet confirmed if these specific prompt updates apply to other models. ## Balancing Innovation and Compliance Anthropic intends to continue iterating on these prompts to strike an appropriate balance between utility and safety. Addy Osmani invited users to report instances where Claude incorrectly rejects legitimate, original projects. This feedback loop is essential for refining the AI's ability to distinguish between creative inspiration and direct copyright infringement. ## Context of Retro Gaming The gaming industry is currently seeing a surge in AI-assisted reverse engineering, including decompilations and web ports. These techniques often rely on automated tools to modernize or port legacy software. Anthropic’s intervention highlights the tension between the accessibility of AI coding tools and the legal protections surrounding classic software titles. ## Operational Impact on Developers Developers using Claude Code for game design must now ensure their project documentation clearly reflects original intent. Since the AI now checks for specific title recreations, users working on derivative projects may face increased friction. This requires developers to be more transparent about their design assets and project goals during interaction. ## Future of AI Coding Tools This move signals a broader trend where AI providers implement guardrails to mitigate legal exposure for their platforms. As AI coding assistants become more capable of complex software generation, defining the boundaries of fair use will remain a critical challenge. Future updates will likely focus on more nuanced detection of IP. ## Sources covering this story - https://www.theverge.com/ai-artificial-intelligence/1024632/anthropic-claude-ai-video-game-clones — --- _Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._