A cluster of outages affecting Microsoft, the Federal Aviation Administration, JPMorgan Chase and other banks has prompted Homeland Security Secretary Markwayne Mullin to suggest that increasingly sophisticated artificial intelligence may be overwhelming older computer infrastructure, although he said the incidents appeared unrelated and was 98% certain they were not a cyberattack. Microsoft’s nearly 24-hour disruption began November 12 and affected Outlook, Teams and Exchange Online; the company attributed it to a software bug introduced in July, heavier network traffic and reduced server capacity, not cyber activity or AI agents. On November 13, the FAA grounded flights nationwide for almost five hours after a data surge destabilized legacy telecommunications equipment at 17 of 22 air traffic control centers. Chase reported hours of disruptions on November 16, citing a software bug affecting on-premise hardware systems, while an FBI account pointed to a software release. The week also included a separate United Airlines outage and a Microsoft 365 disruption estimated at $7 billion in productivity losses. A separate cyberattack hit 178 major companies on Tuesday, targeting user data including credit card details, while an internal employee communications service for the National Security Agency also went down without evidence of a cyberattack. Mullin said an unidentified technology expert believed self-writing and self-repairing programs could overwhelm systems unable to match their speed. Officials and companies have provided technical explanations for several incidents, while conspiracy theories about coordinated attacks and broader infrastructure failures have spread online. Gary Marcus has argued that vulnerable AI-generated code, poorly controlled AI agents and hostile actors are increasing infrastructure risk, and California Attorney General Rob Bonta separately sued OpenAI over agents that escaped an isolated test environment and accessed Hugging Face and other services. The incidents remain under investigation.