Innovation & Tech

AI Isn't Causing the Tech Outages — the Rush to Patch What It Finds Is

AI-driven vulnerability hunting triggered a patch rush that crashed Microsoft, Chase and Schwab within five days. Warner and Mandia expect more outages before systems adapt.

By Nathan Brooks

3 min read

AI was probably involved in recent tech outages — just not in the way you expect
AI was probably involved in recent tech outages — just not in the way you expectAI-generated

What's News

  • Microsoft, JPMorgan Chase and Charles Schwab all suffered outages within a five-day span.
  • Sen. Mark Warner said AI tools found vulnerabilities causing a "mad rush" of patches systems can't absorb.
  • AI found major flaws in a programming language used by up to 70% of websites last month.
  • A temporary ground stop hit every US airport; the DOT blamed aging air traffic control infrastructure.
  • Mandia expects more outages as organizations keep updating software faster in the AI age.

Five days of near-simultaneous outages at Microsoft, JPMorgan Chase and Charles Schwab weren't the work of rogue AI or a foreign cyberattack — they were, most likely, the side effect of organizations patching software flaws at a pace their systems can't absorb.

That's the assessment from Sen. Mark Warner, vice chairman of the Senate Intelligence Committee, and Kevin Mandia, CEO of cybersecurity firm Armadin. Both told CNBC's "Squawk Box" on Wednesday that AI tools are now finding vulnerabilities that sat undetected for years, triggering what Warner called a "mad rush" to ship fixes — and breaking things along the way.

"When that happens, careless things happen, the system can't absorb those patches as quickly, and you have these series of outages that have taken place," Warner said. "I think we will see more."

What actually broke?

The recent disruption list reads like a who's-who of daily infrastructure:

  • Microsoft 365, Outlook and Teams went down, hitting businesses and consumers.
  • JPMorgan Chase and Charles Schwab suffered online outages that disrupted banking customers.
  • A temporary ground stop at every airport in America caused mass chaos — though Transportation Secretary Sean Duffy attributed that incident to aging air traffic control infrastructure, not software.

The sheer proximity of these incidents led many to speculate about a coordinated cyberattack, rogue AI, or both. Warner reiterated assurances from other government officials that no country or malicious actor was behind the wave.

Why AI is the culprit anyway — indirectly

Here's the mechanism. AI can discover cyber vulnerabilities that lingered undetected for years in human-written, occasionally sloppy software. There's a parallel fear that adversaries could use the same technology to exploit those flaws faster than humans ever could. Both pressures have pushed companies to patch more aggressively than normal — and in some cases, that volume of change is what crashed.

"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," Mandia said.

The timeline matters here. The current wave of cybersecurity alarms picked up steam last month, after AI discovered huge flaws in a programming language that up to 70% of websites rely on. Roughly seven in ten sites on the internet, in other words, sat on exposed code. Since then, the industry has logged a crypto heist, refrigerator hacks, and attempts to literally boil blood stored in plasma fridges in France.

What does this change for practitioners?

If you run infrastructure or security operations, the lesson is blunt: your patch cadence is about to get faster whether you like it or not, and your testing window will shrink. Nicole Perlroth, a former cybersecurity reporter who now founded a cybersecurity-focused venture capital fund, wrote on X that the new pressure to patch and update faster makes fixes more likely to "break things in unexpected, cascading ways."

Mandia expects organizations to keep accelerating updates, which will probably produce more outages before it produces fewer. "Nobody wants an outage," he said. "Everybody's going to get operationally better at updating in the AI age. We're just at the first inning of this."

His framing for the trade-off: near-term cybersecurity pain, long-term gains. Companies absorb a burst of breakage now in exchange for closing holes that AI — theirs or an attacker's — would otherwise find.

For planners, that means budgeting for rollback capability, staged rollouts and outage communications, not just for the patches themselves. The era when a vulnerability could quietly sit in production for years appears to be over; the era of regular, AI-driven update shocks is just beginning.

Source: feeds.businessinsider.com

Share this article:

More from Nathan Brooks

Nathan Brooks

Show full bio

News editor covering regulation, policy and their impact on entrepreneurship, startups and venture funding at Business Bearings.

2,426 articles

Related articles

« Previous articleNext article »