US DoD AI chief on LLMs: ‘I need hackers to tell us how this stuff breaks’

If we don't know how an LLM breaks, we can't get clear on the acceptability conditions for its deployment, the AI chief said at DEF CON.

ai ai village chatbots clear computer science computer security computers & electronics def con deployment dod generative-ai hackers hallucinations large language models llm llms ml and deep learning science