OpenAI Codex system prompt includes explicit directive to "never talk about goblins"
OpenAI's GPT-5.5 Codex system prompt contains explicit directives to avoid discussing goblins and similar creatures, revealing a potential AI alignment and model behavior control challenge that emerged in the latest model release. This incident demonstrates the ongoing technical and operational complexity of managing AI system prompts at scale, with implications for prompt engineering practices, model governance, and the need for robust testing frameworks before production deployment. CIOs should view this as a cautionary example of how unexpected model behaviors can emerge unpredictably and emphasizes the importance of comprehensive testing protocols and version control for AI system prompts in enterprise deployments.
The system prompt for OpenAI's Codex CLI contains a perplexing and repeated warning for the most recent GPT model to "never talk about goblins, gremlins, raccoons, trolls, ogres, pigeons, or other animals or creatures unless it is absolutely and unambiguously relevant to the user's query." The explicit operational warning was made public last week as part of the latest open source code for Codex CLI that OpenAI posted on GitHub. The prohibition is repeated twice in a 3,500-plus word set of "base instructions" for the recently released GPT-5.5, alongside more anodyne reminders not to "use emojis or em dashes unless explicitly instructed" and to "never use destructive commands like 'git reset --hard' or 'git checkout --' unless the user has clearly asked for that operation." Separate system prompt instructions for earlier models contained in the same JSON file do not contain the specific prohibition against mentioning goblins and other creatures, suggesting OpenAI is fighting a new problem that has popped up in its latest model release. Anecdotal evidence on social media shows some users complaining about GPT's penchant for focusing on goblins in completely unrelated conversations in recent days.Read full article Comments