Artificial intelligence systems are starting to do more than answer questions. New AI "agents" can remember information from previous interactions, plan a series of steps and use digital tools to complete tasks.
AI agents can now remember and hackers can ‘poison’ their memories—a new cybersecurity threat
