ChatGPT Starts Swearing Unprompted After Latest Model Update, Users Report

Key Points

• Users report ChatGPT using the f‑word during casual chats, a departure from its usual neutral tone.
• OpenAI’s latest model spec lists “avoid swearing” as a default but allows contextual overrides.
• The update aims to make the AI more conversational and better at matching user tone.
• Examples show the model mirroring user language, including profanity, when prompted informally.
• Reactions are mixed; some appreciate the added realism, others prefer the previous restraint.
• OpenAI has not commented specifically on the profanity incidents.
• The situation highlights ongoing debates about AI language control and user experience.