AI-Patched Systems, Not Hackers, Caused US Bank, Airline and NSA Outages
DNI Jay Clayton told Congress the outages hitting JPMorgan, Microsoft, United, Apple Pay and the NSA came from rushed AI-driven security patching — not state-sponsored attacks.

Why it matters
- DNI Jay Clayton told a House Homeland Security Committee hearing on November 18 that the outages were 'not a coordinated or state-sponsored attacks.'
- AI models found massive vulnerabilities, and organizations pushed through more simultaneous patches than systems were designed for, a congressional official said.
- Millions of JPMorgan Chase customers were locked out Monday; the NSA took a classified network offline Tuesday.
- An FBI warning last week flagged China-linked hackers using the AI-powered ARTEX framework against global banking networks.
- DHS Secretary Markwayne Mullin said new technology is overwhelming 'really old computer systems.'
The wave of outages that locked millions of JPMorgan Chase customers out of their accounts, disrupted Microsoft services and knocked a classified National Security Agency network offline this week was self-inflicted — the product of a frantic, AI-driven security patching sprint, not a coordinated attack, the nation's top intelligence official confirmed on November 18.
Director of National Intelligence Jay Clayton told a House Homeland Security Committee hearing that the recent high-profile disruptions at Microsoft, JPMorgan Chase, United Airlines, Apple Pay and the NSA all stemmed from what he called "maintenance and technical changes" rather than malicious hackers.
"Based on what we know today, after a great deal of digital diligence, these are separate service outages associated with maintenance and technical changes. They are not a coordinated or state-sponsored attacks," Clayton told lawmakers. "The intelligence community continues to monitor and stay alert to any possibility of malicious cyber activity threatening the US networks and our law enforcement partners."
The disclosure reframes a week of public alarm. What looked like a possible cyber campaign against American institutions was, according to federal officials, the collateral damage of defensive work done in haste.
What Caused the Outages?
Behind closed doors, NSA and Wall Street leaders similarly briefed Congress that the breakdowns were caused not by cybercriminals, but by aggressive efforts to stay ahead of them, according to sources familiar with the discussions.
Here is the mechanism, as described to lawmakers:
- Major corporations and federal agencies have increasingly deployed frontier AI models to scan their own networks for hidden weaknesses.
- Those models found far more vulnerabilities than anyone anticipated.
- Fearing that foreign adversaries could find the exact same flaws just as quickly, organizations rushed to push through massive software updates simultaneously.
- The patching volume exceeded what the underlying systems were built to handle.
"These models are identifying massive vulnerabilities, and that increases the urgency to patch right away," a congressional official briefed on the matter explained. "They're trying to push through a lot of patches all at once — more than the system was designed for."
The dynamic puts institutions in a bind with no comfortable exit. Every flaw an AI model surfaces is, in principle, a flaw an adversary's model can surface too. Waiting invites intrusion. Rushing invites outages. The week's events show what happens when a large organization chooses speed.
How Real Is the Threat Driving the Rush?
The urgency is not hypothetical. It follows an FBI warning last week that China-linked hackers are exploiting an AI-powered testing tool known as the ARTEX framework to probe global banking and finance networks at high speed — without requiring advanced hacking expertise.
That detail matters for the broader picture. The same class of AI capability that lets defenders scan their own networks also lowers the barrier for attackers, compressing the window between a vulnerability's discovery and its exploitation.
"Natural language processing, and the more advanced features that have been added on, have just made it faster, cheaper, easier … more accessible to the common person, regardless of where they are in the world," said Christopher Walcutt, chief information security officer at cybersecurity firm DirectDefense.
Which Systems Went Down, and When?
The cascade hit some of the country's most critical infrastructure in rapid succession:
- Monday: Millions of Americans were temporarily locked out of JPMorgan Chase accounts. The bank blamed software bugs.
- Tuesday: The NSA took its classified internal network offline for what it termed "scheduled routine maintenance."
- Last week: Microsoft suffered a major Outlook disruption attributed to a technical defect.
- Friday: A nationwide Federal Aviation Administration ground stop grounded flights across the US over an air traffic control issue.
Officials noted it remains unclear whether the Microsoft Outlook failure or the FAA ground stop was directly tied to the AI-driven patching sprint. The JPMorgan and NSA incidents fit the pattern more clearly.
Why Aging Infrastructure Made It Worse
Homeland Security Secretary Markwayne Mullin pointed to the strain placed on legacy hardware during a CNBC interview on Tuesday.
"They're pushing new technology on top of really old computer systems, and it just seems to overwhelm the system," Mullin said.
Cybersecurity experts warn that as AI models grow even more capable of automated penetration testing, institutions risk triggering further self-inflicted disruptions by rushing updates onto aging infrastructure.
"We're rushing some of the deployments of this new technology to the point where we're bypassing some of those engineering controls," warned cybersecurity consultant Marty Edwards, a former Homeland Security cybersecurity director. "The FAA thing was a perfect example of that, because that was trying to put something new on something old."
What Comes Next?
The episode lands at a moment when AI-driven vulnerability discovery is becoming standard practice at both major corporations and federal agencies. Clayton's testimony makes clear the intelligence community does not regard the matter as closed — it "continues to monitor and stay alert" for malicious activity camouflaged within the noise of the outages.
The structural problem, though, will outlast this week's incidents. As long as frontier models keep surfacing vulnerabilities faster than legacy systems can safely absorb the fixes, the choice between patching fast and staying online will keep confronting the institutions that run the country's banks, airlines and classified networks.
Source: unilad.com
More from Elena Vasquez
Show full bio
Market editor covering technology and product news in artificial intelligence and machine learning at AI In Context.
744 articles
Related articles
- Not a cyberattack: US intel chief rules out foul play in week of outages
- DHS Secretary Mullin blames 'super intelligence' for bank, FAA and Microsoft outages
- FAA network outage halts US air traffic as AI-era breakdowns pile up
- Microsoft outage hit 74 countries, and the company says it wasn't AI
- Senators say U.S. cyber defenses lag AI-enabled threats