OpenAI Bans Goblins! Codex AI Scrubbed of Mythical Creature Mentions

Top Summary

• What happened: OpenAI has implemented strict behavioural rules in its Codex AI agent, specifically banning mentions of goblins and other mythical creatures.
• Why it matters: This highlights the ongoing efforts to control the output and behavior of AI models, raising questions about censorship and bias.
• What changes for people: Users of the Codex CLI will no longer be able to generate or discuss content related to goblins or similar mythical entities.
• Who is affected: Developers and users of OpenAI's Codex, particularly those using GPT-5.5, are directly impacted by these restrictions.

Codex's New Restrictions

OpenAI has placed a ban on its AI Agent, Codex, from referencing goblins and other mythical creatures.

Codex is OpenAI's answer to Anthropic's Claude Code AI Agent.

GPT-5.5 Instructions

The newest version has strict behavioural rules.

According to newly released base instructions for GPT-5.5 in OpenAI's Codex CLI, the model is explicitly told to "never talk about goblins.

What to Watch Next

The industry will be watching to see how these restrictions impact Codex's functionality and user adoption. Future iterations of AI models may continue to implement similar behavioral guidelines as AI companies attempt to control outputs.