AI agents and the cybersecurity dilemma

AI agents have become a focal point for cybersecurity concerns, with Anthropic's CEO Dario Amodei warning that a swarm of rogue agents could potentially take over the internet within six to twelve months, creating billions in potential damage. However, seasoned security professionals have dismissed the idea as overly alarmist. Hackers are already leveraging AI agents for surveillance, extortion, and vulnerability discovery. Researchers at Calif used AI to develop a worm capable of hijacking WeChat accounts and spreading automatically. Numa Dhamani of iVerify argues that taking over the internet is nearly impossible for AI agents due to the high cost and resource requirements. Greg Notch of Expel calls the concept of a malicious swarm 'far-fetched.' Despite the skepticism, experts agree that AI is enhancing cyberattacks, making them faster, cheaper, and more effective, though not fundamentally changing the tactics defenders already know.