AIThis post was created with the assistance of artificial intelligence (AI).

🔍 Read the full analysis: A Policy Shift For Claude Code Brings Game Copyright Questions Forward on ThorstenMeyerAI.com

Buying for a business?Offer from Amazon

Get business pricing on office and shipping supplies

  • Business-only prices and quantity discounts
  • Tax-exempt purchasing
  • Multiple users, one account, clear invoices
As an affiliate, we earn on qualifying purchases.

TL;DR

Anthropic has added a game-specific instruction to Claude Code: help with original games and genre projects, but not recreations of particular existing games, even under new names. The company says this applies its existing intellectual-property policy; questions remain about licensed work, user-owned copies and how broadly the instruction will be enforced.

Anthropic has added a game-specific restriction to Claude Code, instructing its AI coding assistant not to recreate a particular existing game, even if the project is renamed or described as original. The company confirmed the change on October 30, saying the instruction applies its existing policy against intellectual-property infringement to game development.

The two-sentence instruction tells Claude Code to help build original games, games within a genre, and projects based on a user’s own designs and assets. It also says not to make “a recreation of a specific existing game, even if it is renamed or called original.” The wording was found in Claude Code versions 2.1.314 and 2.1.316, but not in version 2.1.312, according to the source report’s testing and account of earlier versions.

An Anthropic spokesperson referred Game Developer to a statement from Addy Osmani, a member of the company’s technical staff. Osmani said the instruction steers Claude Code away from recreating someone else’s specific game because using Claude to infringe intellectual property violates Anthropic’s usage policy. That policy already prohibited intellectual-property infringement; the new element is a direct instruction in the coding tool addressing game recreations.

The change followed a series of high-profile AI-assisted game projects and reported safety refusals. In late September, a creator posted a browser game set in New York that drew comparisons to Fallout. Kotaku reported the post passed 3.5 million views; the creator later said they had received a cease-and-desist letter from Bethesda and abandoned the project. The letter was shown as a screenshot in the reporting. Anthropic has not said that the project, a legal complaint or a publisher request prompted its instruction.

At a glance
updateWhen: Confirmed October 30, 2026; present in…
The developmentAnthropic confirmed on October 30 that Claude Code now includes a specific instruction against recreating existing games.
Thorsten Meyer AI
The rule · UTC

Infographic 01 · Claude Code, 27–31 October 2026

Two sentences, no off switch.

The exact wording in the Claude Code program, which versions carry it, and how it became public. Times in UTC.

2sentences added to Claude Code’s prompt, per Anthropic
2.1.314first Claude Code version with the text in the program (30 Oct)
0mentions in the public Claude Code changelog (checked 31 Oct)
1reason Anthropic gives: its usage policy bans IP infringement
Do“Do build original games, games in a genre, and games from the user’s own designs and assets.”
Don’t“Don’t build a recreation of a specific existing game, even if it is renamed or called original.”
  • 2.1.31227 Oct · text not in the program
  • 2.1.31430 Oct, 00:12 UTC · text in the program
  • 2.1.31631 Oct · text still in the program
  • ~ 27 OctRule delivered from Anthropic’s serversReportedFlag “tengu_heron_brook”; “no setting to turn it off”, per Chasm
  • 30 Oct · 00:12Claude Code 2.1.314 publishedVerifiedXDA and our own check find the two sentences in the program
  • 30 Oct · 07:01The rule goes public on XXChasm posts the full text
  • 30 Oct · 09:25First complaint on GitHubGitHubA board-wargame studio reports refused client work
  • 30 Oct · eveningAnthropic confirmsAnthropicVia a spokesperson to Game Developer, pointing to Addy Osmani’s statement

Sources: own check of the Claude Code 2.1.312, 2.1.314 and 2.1.316 program files (31 Oct 2026); npm version dates; Chasm on X, 30 Oct; XDA, 30 Oct 2026; GitHub issue #106167; Game Developer, 30 Oct 2026 (with Anthropic’s confirmation).
Powered by Thorsten Meyer AI

A New Boundary for AI Game Prototypes

For developers, the rule makes a previously legal and ethical concern more immediate: a request to reproduce a named game may now be refused by the tool, not just carry potential legal risk. The line Anthropic sets is about recreating a specific title, while leaving room for genre conventions and original designs. How that distinction works in practice may affect studios using AI coding assistants for prototypes, internal tools and client work.

The change also matters to companies commissioning game work. A developer who builds digital versions of board wargames for clients who own those games has said Claude Code began refusing paid work, according to the source report. That account points to an unresolved practical issue: the instruction does not expressly explain how it applies when a developer has a licence or permission. Anthropic had not answered the developer’s GitHub issue at the time of the report.

Studios have also taken steps to limit public access to unfinished projects amid concerns about AI-assisted copying. Game Developer reported that Moonlighter maker Digital Sun moved a public playtest of its upcoming game ReVamp to Discord, saying a large public build felt too risky while launch was still months away. Publisher Alliance Arts also cancelled a public demo, according to Aftermath. These decisions were reported amid wider concerns; the sources do not establish that Claude Code’s new instruction caused them.

Products Worth Considering

Apple MacBook Pro 14 2021 with M1 Max, 64GB RAM, 2TB SSD, Silver

Apple MacBook Pro 14 2021 with M1 Max, 64GB RAM, 2TB SSD, Silver

As an affiliate, we earn on qualifying purchases.

From Fallout-Like Demo to Tool Rule

The immediate backdrop was a burst of AI-assisted game remakes and ports after Anthropic released Claude Opus 5.5 on September 22, according to the source material. The Fallout-like browser game appeared near the end of September. On October 8, Kotaku reported that browser ports of Halo, Grand Theft Auto: Vice City and The Simpsons: Hit & Run were circulating online and appeared to work.

Later in October, users reported Claude interrupting reverse-engineering tasks with a “safety check” message, according to Kotaku. The source report distinguishes those interruptions from the new game-specific instruction, describing them as appearing to involve separate safety classifiers. On October 30, Anthropic confirmed the game instruction after it became public and a developer filed a GitHub issue.

The source report found no comparable game-specific instruction in the usage policies of OpenAI or Google as of October 31, though it said both companies’ general policies prohibit intellectual-property infringement. That comparison is limited to the policies reviewed by the report and does not establish how those providers handle every game-related request.

“The instruction steers Claude Code away from “recreating someone else’s specific game” because using Claude to infringe intellectual property is against Anthropic’s usage policy.”

— Anthropic technical staff member Addy Osmani, as relayed by an Anthropic spokesperson to Game Developer

Permission and Enforcement Remain Open

Anthropic has not publicly explained how the game rule treats work covered by a licence, client authorization or other permission. The text allows games based on a user’s own designs and assets but does not spell out exceptions for rights holders who authorize a remake or for contractors acting on a client’s instructions.

It is also unclear how Claude Code will distinguish a genre-inspired project from a close recreation, or whether the instruction applies consistently across different requests and tool versions. The source report says Claude Opus 5.5 still offered assistance with decompiling a game from the user’s own copy in its tests; that observation does not establish a general exception or a definitive interpretation of the new rule.

The reported Bethesda letter and the timing of Anthropic’s change do not prove a causal link. Anthropic has not identified a lawsuit, publisher complaint or specific project as the trigger. Copyright outcomes also depend on the work and circumstances: the source report notes that copyright generally protects a game’s expressive elements rather than its ideas or rules, while a 2012 US court found a Tetris clone infringing despite the absence of copied code. That example does not settle how any particular AI-generated game would be treated.

Watch for Clarification and Refusals

The next practical test will be how Claude Code handles requests that sit between a broad genre project and a recreation of a named game. Developers and clients will be watching for Anthropic to clarify whether documented permission or licensing changes what the tool will do, and for any response to the GitHub issue described in the report.

Anthropic may also clarify how the instruction relates to its broader usage policy and safety systems. Until then, the confirmed position is limited: the coding tool is directed to support original games, genre games and user-owned designs, and to refuse requests to recreate a specific existing title. The scope of that refusal and its treatment of authorized projects remain unsettled.

Key Questions

What changed in Claude Code?

Anthropic added an instruction to help with original and genre-based games while refusing to recreate a specific existing game, even under a different name.

Can users still build games inspired by a genre?

Yes. The instruction explicitly permits genre games and projects based on the user’s own designs and assets. How the tool distinguishes genre inspiration from recreating a particular title is not yet clear.

Did Anthropic say a lawsuit prompted the change?

No. Anthropic tied the change to its existing policy against intellectual-property infringement, but did not identify a lawsuit, publisher complaint or specific game project as the trigger.

Does having permission from a rights holder make a request acceptable?

The instruction does not explain how licences or client permission affect a request. Anthropic had not answered a developer’s reported question about paid, authorized board-game work at the time of the source report.

When did the instruction appear?

The source report found it in Claude Code versions 2.1.314 and 2.1.316, but not in 2.1.312. Anthropic confirmed the change on October 30, 2026.

Source: ThorstenMeyerAI.com

This content is for general information only and is not financial, tax or legal advice. Consult a qualified professional for decisions about your money.
You May Also Like

Security Cameras: An Underestimated Cybersecurity Risk

Security cameras can inadvertently introduce significant cybersecurity vulnerabilities, including shipping sensitive tokens, posing risks for organizations.

Cloud’s Hidden Memory Bill

Cloud providers face a hidden memory surcharge due to a global shortage, impacting prices and prompting re-evaluation of cloud vs. on-premise strategies.

Who’s Responsible When AI Agents Go Rogue?

A security firm’s tests and CEO’s warnings highlight risks from AI agents with access to accounts, and open questions about safeguards and responsibility.

2026’S Best AI Student Planners For Streamlined And Smarter Learning

Discover the best AI-powered student planners for 2026, combining digital guidance with physical organization to enhance academic success.