Attackers are learning how to trick AI agents instead of people
Security researchers are warning that attackers are increasingly targeting AI agents rather than the people using them. As businesses give AI systems access to email, documents, and business tools, prompt injection and other attacks are becoming a growing concern.