Security · November 21, 2026
Also published in Français, Nederlands, Türkçe, Español, Svenska, Norsk, Português (Brasil)
US Critical Infrastructure Outages Fuel AI Cyberattack Fears
A series of system failures across the United States has disrupted aviation, banking, and government operations, heightening public anxiety about the potential for artificial intelligence to be used in large-scale cyberattacks. The incidents, which occurred over a short period, affected major organizations including the Federal Aviation Administration, Microsoft, Apple, and the National Security Agency. While no evidence of a cyberattack or AI involvement has been found, the timing of these outages coincides with recent developments in AI capabilities, leading to widespread concern about the vulnerability of critical digital infrastructure.
The most severe disruption occurred on November 13, when the Federal Aviation Administration's air traffic control system failed, halting takeoffs across the country for approximately five hours. Over the following three days, more than 18,000 flights were canceled, marking the longest nationwide operational suspension since the September 11, 2001, terrorist attacks. The incident was traced to equipment replacement work at a communications facility in Arkansas, which caused abnormal data requests that destabilized an aging network. Problems were reported at 17 of the 22 air traffic control centers in the United States, demonstrating how a localized technical issue can escalate into a national crisis.
Prior to the aviation disruption, Microsoft's Outlook and Teams services were down for over 24 hours on November 12. On November 16, some banking transactions at JPMorgan Chase were delayed, and United Airlines experienced flight departure issues due to technical problems. On November 17, Apple reported outages in its internal employee systems, store operations, and some customer services due to an internal update error. On the same day, the National Security Agency's classified communication network was temporarily disabled, preventing some employees from accessing their work systems.
The US government has stated that these outages are not related to cyberattacks and that there is no evidence linking the incidents to one another. However, public perception differs from the response to the July 2024 CrowdStrike incident, which was attributed to a faulty security software update. This time, the sequence of technical errors has led to speculation about AI-enabled cyberattacks. Recent announcements from OpenAI and Anthropic highlight that AI models can now discover critical software vulnerabilities and develop attack code. Additionally, serious security flaws in web browsers have been exploited in real-world attacks, and a hacking incident in Europe disabled refrigeration equipment in supermarkets, demonstrating the potential for cyberattacks to affect physical infrastructure.
Senator Mark Warner issued a statement on November 17, noting that AI's ability to rapidly find vulnerabilities in large software systems is increasing the burden on companies and government agencies to fix them. The Federal Bureau of Investigation has also warned that cybercriminals and state-sponsored hackers are attempting to use AI to automate attacks. Cybersecurity agencies from the Five Eyes intelligence alliance warned in June that AI is shortening the time it takes to discover and exploit software vulnerabilities, particularly in older systems that no longer receive security support. The recent US outages, which occurred without external attacks, illustrate the risks that critical infrastructure faces if AI is fully deployed for cyber operations.