Table of Contents
Recent widespread technical failures, ranging from email outages to air traffic control issues, highlight a growing fragility in the global digital infrastructure. Cybersecurity experts suggest that while these incidents are often unrelated, they are occurring in an environment increasingly destabilized by the rapid advancement of artificial intelligence.
The Rise of the Vulnpocalypse
Adam Meyers, senior vice president of counter adversary operations at CrowdStrike, described the recent surge in digital instability as a trend that is unfortunately not improving. A primary driver of this volatility is the way AI has transformed cybersecurity, a phenomenon some experts have dubbed the “vulnpocalypse.”
Previously, executing a successful cyberattack required significant skill and resources. Today, AI models have effectively democratized hacking, allowing individuals with limited expertise to identify and exploit vulnerabilities. These AI tools can scan code and websites for flaws much faster and more cheaply than humans, exposing a massive backlog of security issues. Consequently, the time between the discovery of a vulnerability and its exploitation has shrunk from weeks to mere hours.
The Vibe-Hacker Phenomenon
John Strand, CEO of Black Hills Information Security, noted that AI has automated the tedious “grunt work” of offensive security. Attackers are now using sophisticated AI harnesses—tools that direct models to perform specific tasks—to refine their efforts. One notable example is Artex, a tool that recently prompted an FBI warning. According to Meyers, CrowdStrike identified a 26-year-old attacker in China who used Artex to hack South Korean banks and subsequently used AI to facilitate the sale of stolen data.
While major AI developers have attempted to place guardrails on their most powerful models, experts warn that open-weight models, which can be run on private hardware, remain beyond the reach of corporate or government oversight. “The genie’s out of the bottle, and there’s no putting it back in,” said Gary Barlet, public sector chief technology officer at Illumio.
The Patching Race
Defenders are attempting to counter these threats by using the same AI tools to identify and patch vulnerabilities before they are exploited. Programs like OpenAI’s Daybreak, Anthropic’s Project Glasswing, and Google’s Fairwind provide security experts with early access to powerful models. However, the sheer volume of disclosures is overwhelming; the National Vulnerability Database recorded over 10,000 vulnerabilities in November alone, compared to fewer than 5,000 in October 2025.
Microsoft CEO Satya Nadella noted at the New York Times Dealbook Summit that the speed of modern exploits leaves little room for testing. “Right now the patch needs to go instantly from being available to being patched in production because it’s being exploited in the wild,” Nadella said. This urgency is contributing to system instability, as rushed updates can sometimes cause the very outages they are intended to prevent.
AI-Generated Code and Workforce Shifts
The reliance on AI to write software is also creating new risks. As companies reduce human engineering teams to cut costs, AI is being used to generate more code with less human oversight. This practice risks introducing new vulnerabilities into the software supply chain. Ted Miracco, CEO of Approov, described the task of cleaning up these systemic issues as “daunting.”
Despite the current chaos, some experts remain optimistic about the long-term outlook. Matt Wyckhouse, CEO of Finite State, expects the next two years to be a period of difficult “growing pains” as the industry works through the backlog of vulnerabilities. He predicts that software will be significantly more secure once these issues are addressed.
In the meantime, experts emphasize that both individuals and corporations must prioritize regular software updates. Barlet also suggested that consumers should demand higher standards of security and privacy from the companies they rely on, rather than accepting frequent data breaches as a normal part of digital life.








