So normally with Copilot it will generate three suggested prompts that will continue the conversation, e.g. “Tell me more about blank”, “Are blank healthy?” and “Where can I find blank nearby?”

Here, the autosuggested responses were drafted from the perspective of Copilot itself, not from that of the user, responding to my message. It’s like, Copilot suggesting a response apologising to itself.

I really hope this is evidence of LLM CJD.