[tabletcorry]

Your added paragraph appears to suggest the opposite, that this was an LLM response. Was the "leaked data" a response from an LLM directly?

[JyB]

Yes apparently which makes this report pretty flimsy.

[tptacek]

Upthread, OpenAI's security team confirms it's a false report; it's a variant of the empty-prompt hallucination.