Generative AIs Biggest Security Flaw Is Not Easy to Fix

Chatbots like Open AIs ChatGPT and Googles Bard are vulnerable to indirect prompt injection attacks. Security researchers say the holes can be pluggedsort of.