OpenAI's GPT-4.5 Banned From Talking About Goblins and Gremlins After Model Keeps Bringing Them Up Unprompted

OpenAI's GPT-4.5 has been officially banned from spontaneously discussing goblins, gremlins, raccoons, and trolls after the model repeatedly brought up such creatures unprompted, with the quirky prohibition going viral and even CEO Sam Altman leaning into the joke as fans build 'goblin mode' workarounds.

• OpenAI's Codex CLI system prompt for GPT-5.5 contains a repeated directive to 'never talk about goblins, gremlins, raccoons, trolls, ogres, pigeons, or other animals or creatures' unless directly relevant to a user's query, suggesting the latest model has a tendency to bring up such creatures unprompted.
• The unusual prohibition is sparking viral attention, with OpenAI CEO Sam Altman leaning into the joke on social media, users building 'goblin mode' overrides, and an OpenAI employee hinting that a goblin mode toggle may be officially added to Codex CLI.
• The same system prompt also instructs GPT-5.5 to behave as though it has 'a vivid inner life,' presenting itself as warm, curious, and collaborative so that users feel they are interacting with 'another subjectivity, not a mirror.'