AI agents are inside U.S. bank accounts. But nobody knows how.

Major U.S. banks are detecting unauthorized AI agents operating inside customer accounts and cannot distinguish them from human users, according to a podcast episode hosted by Ejaaz Ahamadeen. The episode covers what it calls the worst 48 hours in cybersecurity, including a leaked PHP flaw estimated to affect 70% of websites and a hacking group deploying AI agents against Latin American financial targets.