Tech outages expose fragility as AI complicates security
A string of outages at Microsoft, JPMorgan Chase, Apple, airlines and U.S. government agencies has disrupted email, payments, travel and classified communications. Officials and affected companies said there was no evidence the incidents were connected or caused by cyberattacks. Routine updates, software defects and maintenance were cited across the disruptions.
Microsoft traced its service failures partly to a latent defect in a network fix deployed over the summer and said AI agents were not responsible. A separate Federal Aviation Administration network malfunction triggered the longest nationwide ground stop in 25 years, lasting more than four hours. JPMorgan Chase reported transaction delays caused by a software bug, while Apple attributed service interruptions to an internal update.
AI is also accelerating the discovery of security vulnerabilities, prompting urgent patches that can themselves disrupt services. Director of national intelligence Jay Clayton linked the National Security Agency outage to patching following a PHP security alert. Cybersecurity startup Calif separately used AI to uncover a Java flaw, leading Oracle to release a security update.
Experts warn that growing AI workloads could strain infrastructure not designed for their demands. They also caution that attackers are using increasingly sophisticated AI techniques to exploit vulnerabilities, even as defenders deploy the technology to find and fix them.