Would you like to get Dart Vader to give you bomb-making recipes? Now you can, using #LLM jailbreaking exploits that point to significant structural and unaddressed vulnerabilities in the technology, explains David Kuszmar
spectrum.ieee.org/jailbreaking... #AI
spectrum.ieee.org
How Darth Vader Taught Me Card Counting and AI Security Got Weird
It only took a little prompting to hijack the biggest AI models