AI Researchers Say They've Invented Incantations Too Dangerous to Release to the Public

A team of researchers found prompts that are so effective at tricking AI models that they're keeping them under wraps.

Artificial Intelligence https://ift.tt/sRY8kyi
AI Transformations