Recent tech outages at Microsoft and Chase have raised questions about the potential role of AI in these incidents. Organizations have increasingly used AI to identify vulnerabilities, which has led to a surge in software updates and patches. This rapid pace may have overwhelmed systems, according to politicians and tech executives.
Kevin Mandia, CEO of the cybersecurity firm Armadin, stated, “We’re updating to withstand and be more resilient, and sometimes you do that at a speed where you couldn’t test all the potential outcomes of the update.” Over the past week, disruptions at Microsoft, JPMorgan Chase, and Charles Schwab have significantly impacted businesses and consumers. Additionally, a temporary grounding of planes in the U.S. created further chaos. Although the Department of Transportation Secretary blamed outdated infrastructure for the air traffic issue, the timing of these outages has led to speculation about coordinated cyberattacks or rogue AI involvement.
AI has the capability to find long-overlooked cyber vulnerabilities, presenting both a tool for security enhancement and a risk for exploitation. The rapid application of patches to address these vulnerabilities might have inadvertently caused system failures. Mandia emphasized that organizations will likely continue accelerating their patching efforts, which may lead to more outages. He observed, “Nobody wants an outage. Everybody’s going to get operationally better at updating in the AI age. We’re just at the first inning of this.” Although there could be immediate challenges, Mandia suggested that these efforts would lead to stronger cybersecurity in the long term.
Senator Mark Warner, vice chairman of the Senate Intelligence Committee, reassured that recent outages were not due to a cyberattack by another country or malicious actors. Warner highlighted that AI tools identify existing software vulnerabilities, sparking a “mad rush” to implement fixes. This haste can lead to further outages, as systems struggle to integrate these patches swiftly.
Cybersecurity expert Nicole Perlroth noted the pressure to quickly address software issues often results in unexpected complications. An example of the AI-driven identification of vulnerabilities occurred last month, revealing significant flaws in a programming language used by about 70% of websites. This situation has been followed by incidents such as a crypto heist, refrigerator tampering, and extreme cyber threats.
