The Register · Security·2h agoAdd one more AI worry to the nightmare scenario: self-replicating prompt injections#openai#prompt-injection#ai-agents 3 min