OpenAI tells ChatGPT models to stop talking about goblins
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- OpenAI flagged a 175% rise in goblin mentions and a 52% rise in gremlin mentions in GPT‑5.1 outputs after its November launch.
- Codex was instructed to never mention goblins, gremlins, raccoons, trolls, ogres, or pigeons unless absolutely relevant, per OpenAI’s internal code changes.
- ChatGPT’s ‘nerdy personality’ training inadvertently rewarded creature references, prompting a model‑behavior revision.
- OpenAI’s blog post on Thursday disclosed the mitigation steps, denying any marketing motive.
Why it matters: OpenAI trims GPT‑5.1’s ‘nerdy personality’ after a 175% surge in goblin mentions, protecting user experience and brand perception while limiting the model’s spontaneous metaphorical flair.
