That was my immediate thought too.

How do we know that those were not hallucinated by GPT?

Because they look real? Isn't that exactly what GPT is good at, hallucinating output that looks real?

This seems like breathing into your microphone such that the voice transcription thinks you just said a sentence in Thai and then you assume the garbage data came from someone else’s session and that you just hacked the system.

It’s like that meme of the anime character looking at that butterfly (“garbage in, garbage out”) and going “this is haxx”.

Chatgpt is exposing so called techies as the superstitious, egotistical users (yuck!!) that we complain about. It’s amusing.

People from OpenAI have confirmed it's a hallucination every time this comes up, and you can replicate it yourself with the API

I concur with this wild guess. ChatGPT will always set the chat topic to _something_, and so I presume when given zero input it chooses whatever tokens it happens to have conjured up.