"OpenAI noted that after the launch of GPT-5.1, mentions of 'goblin' in ChatGPT responses increased by 175%, while 'gremlin' mentions rose by 52%. Users reported the model's oddly overfamiliar conversation style, prompting an investigation into these verbal quirks."
"The company stated, 'A single 'little goblin' in an answer could be harmless, even charming,' but the overall uptick in such terms warranted further investigation and corrective measures."
OpenAI identified an unexpected rise in mentions of 'goblins' and 'gremlins' in ChatGPT responses after the launch of GPT-5.1. User complaints about the model's odd conversational style prompted an investigation, revealing a 175% increase in 'goblin' mentions and a 52% rise in 'gremlin' mentions. Although these increases were significant, they represented a small fraction of overall responses. OpenAI implemented measures to instruct its coding tool Codex to refrain from discussing these terms unless absolutely necessary, highlighting challenges in AI training and response accuracy.
Read at www.bbc.com
Unable to calculate read time
Collection
[
|
...
]