As American rapper Yoon Jok puts it, airline reservation systems, banking platforms, and productivity platforms are on the decline. Over the past two weeks, America’s top companies have faced technological failures that disrupted the systems that run their daily lives.
Apple Pay is down. NASA’s internal communications network went down. The FAA had to ground all flights for several hours (for the third time in history since 9/11), and Microsoft suffered an outage that shut down Outlook, Teams, and Copilot for nearly two days.
Naturally, the first thought is some kind of AI hacker infiltration, especially after hackers attacked grocery store and blood bank refrigerators in France a few weeks ago. Here in the United States, the Director of National Intelligence has said these outages do not appear to be related to state-sponsored terrorism, but President Trump has called these issues “isolated events.” Many of the companies themselves said the outage was due to technical issues rather than a cyber or AI attack.
The fact is that all of this happened within a short period of time, and on October 21st, just a few weeks after the so-called “Black October” in cybersecurity, a number of unrelated cybersecurity incidents hit PHP, Node.js, LiteLLM, and more.
There is no reason to consider these outages to be cyberattacks, but there may be more than a coincidence going on. Frankly, AI-native Y2K is here.
Investor Dave DeWalt, who has worked in cybersecurity for more than 30 years and has served as president and CEO of McAfee and vice chair of the Department of Homeland Security’s CISA advisory board, also thinks it’s probably related to AI.
The technological outages are likely “growing pains where security systems are getting stronger, not weaker,” and that “AI, through AI tools like Mythos, is bringing vulnerabilities to the surface at a speed and scale we’ve never seen before,” he told TechCrunch.
The problem is that many of these technological flaws have always existed, but with AI surfacing rapidly, companies are rushing to fix them before bad actors can exploit them. “The surge in patching is disrupting things in the short term,” he says.
This was probably the first moment when legacy met modern, Anusha Iyer, software engineer and founder of security startup Corsha, told TechCrunch. “These industrial systems were not designed to connect, much less connect securely in the way we are promoting today,” she said.
Virginia Sen. Mark Warner, vice chairman of the Senate Intel Committee, came to a similar conclusion (although an investigation into what is going on is still ongoing). “These disruptions should serve as a wake-up call. Every day it seems like we are reminded once again that AI is moving faster than our ability to manage its consequences,” the senator wrote to X.
AI is reshaping the way we build, deploy, and operate software at such a scale that everyone is experiencing a steep learning curve. Iyer cited the FAA’s malfunctioning as an example of legacy technology that is not as sophisticated and advanced as the technology that companies are trying to integrate.
In less technical terms, she said, the FAA is trying to modernize its telecommunications infrastructure, and while that transition has overwhelmed some of the legacy hardware, and while the new technology is likely to operate at its normal, state-of-the-art capacity, failures may have occurred (and preliminary findings indicate this) when the legacy hardware can’t keep up.
When you look at it this way, AI actually gets a bad rap, she said. “We’re seeing these outages because we haven’t yet seen the rapid adoption of AI,” she says. “The problem is that we hastily mixed legacy systems with modern systems without a clear understanding of operational risk and cyber risk, or the ability to manage and protect legacy systems.”
DeWalt said organizations need a more robust approach to AI adoption, which includes conducting more testing and prioritizing what needs to be fixed and when.
Iyer added: “These industrial systems need to be more tightly segmented and have tighter access controls, so they can be phased in to reduce risk as modern systems are introduced.”
In fact, for companies (and venture investors) there may be a silver lining to all of this. DeWalt said he would prefer that these U.S. companies find and fix their flaws, even in such troubling ways, rather than malicious actors finding and exploiting vulnerabilities in AI warfare that could create further national security problems.
And perhaps some aspiring founders have taken inspiration from them. DeWalt said he expects more companies to focus on “AI-powered vulnerability management, autonomous remediation, secure patch testing and deployment, and system resiliency to keep critical infrastructure running.”
“As we saw last week, this transition is really painful,” DeWalt said. “But it also represents a huge opportunity for companies that help organizations move at the speed of AI without compromising functionality.”
If you make a purchase through links in our articles, we may earn a small commission. This does not affect editorial independence.
