ChatGPT has been responding in Korean to English-speaking users without any apparent trigger. No language setting change, no Korean text in the conversation - just a mid-reply switch that persists until the user manually tells it to stop.
The reports are consistent: the switch happens inside an existing conversation, sometimes after several English exchanges, and the model continues in Korean until corrected. OpenAI has not acknowledged the issue publicly.
The most likely cause is how ChatGPT infers language. The model doesn't rely solely on your account's language settings - it actively reads the content of your conversation and makes its own inferences. If something in your prompt pattern-matches to Korean-language content in its training data (certain character patterns, topics, or structural cues), it can flip output language without any deliberate trigger on your end.
To fix it: type "please respond in English only" in your next message. For a more reliable reset, start a new conversation and specify the language in your opening message. Account-level language settings in ChatGPT should theoretically prevent this, but they are not consistently overriding the in-context inference for affected users.