Rush to patch bugs found by AI is being blamed for the US tech outages, The Washington Sun reports

AI & Compute

AI-found bugs and rushed patches blamed for wave of US tech outages

Wall Street executives and the NSA privately told Congress that rushed fixes for AI-discovered vulnerabilities caused last week's US outages. Companies tell a different story.

By
Nathan Brooks
Filed
Channel
AI & Compute
Read
4 min read

Wall Street executives and the National Security Agency have told members of Congress that the past week's widespread US technology outages were caused by a rush to patch vulnerabilities identified by artificial intelligence, The Washington Sun reported on Tuesday, citing three sources familiar with the private conversations. The fixes temporarily overwhelmed the systems, those sources said. That claim sits awkwardly beside the public explanations from the companies involved, and nobody in an official capacity has reconciled the two.

A congressional official briefed on the matter described the mechanics bluntly. "These models are identifying massive vulnerabilities, and that increases the urgency to patch right away," the official told the paper. "They're trying to push through a lot of patches all at once, more than the system was designed for."

The Financial Times followed on Wednesday with its own report under the headline "US tech outages blamed on rush to fix bugs found by AI." Senator Mark Warner, vice chair of the Senate Intelligence Committee, told the FT there is no indication the incidents result from a cyberattack, foreign interference or a coordinated effort. "What appears to be happening instead is a consequence of how quickly artificial intelligence is changing the cyber security landscape," he said, as quoted by Gizmodo. A foreign adversary would have been tidier, in a way. Sheer patch volume is harder to deter.

What do the companies themselves say?

The organizations that suffered outages have offered different causes. Microsoft said its November 12 outage was a technical problem, not a cyberattack and not the work of AI agents. JPMorgan Chase blamed a software bug in on-premise hardware. Apple pointed to an internal update. Neither Microsoft nor JPMorgan has publicly tied its outage to patching flaws found by AI, as Gizmodo noted.

The FAA case is stranger. Investigators trace the November 13 ground stop to a newly installed router, and they still do not know why the routers failed. A single router grounding aircraft is a modest failure with a large blast radius.

Where does the AI link stand up?

The strongest official connection is at the NSA. Director of National Intelligence Jay Clayton said on Tuesday that the agency's network disruption came from patching after a PHP vulnerability was found. Reuters reported, according to Gizmodo, that the vulnerability was discovered with AI. The NSA itself described the event as a temporary degradation during routine maintenance, and it did not respond to The Washington Sun's request for comment. One agency, three descriptions. Readers may pick their favorite.

A senior Trump administration official told reporters on Wednesday there is "a great deal of confidence" that the incidents result from system upgrades, patches, hardware changes or other "disparate" events. Asked how far AI was involved, the official conceded it is "challenging to answer that question across all of these incidents that have been reported."

Why has patching sped up?

The pressure on security teams is real and measurable. Sam Rubin of Palo Alto Networks' Unit 42 told CNBC that AI is accelerating vulnerability discovery and forcing IT and security teams to update systems quickly, Newsweek reported.

The October disclosures illustrate the new tempo. Two critical PHP flaws disclosed on October 21 were found with AI tools — one by the security firm Calif and one by Google's Wiz team. The NSA advised at the time that sites unable to patch immediately should go offline. Going offline as the safer option is a remarkable sentence for any infrastructure operator to write down.

The disclosure clock is also tightening. Google's Project Zero said this month that it will cut its disclosure deadline from 90 days to 60 on February 2, 2027, and to 30 on January 3, 2028. OpenJDK and Oracle shipped fixes on November 17 for the Java4Shell flaw.

What changes if the briefing is right?

If the congressional briefing is accurate, the outages mark a structural shift rather than a run of bad luck. AI discovery compresses the window between finding a flaw and exploiting it, which compresses the window for patching, which pushes operations teams toward deploying more changes faster than their systems were designed to absorb. The result is downtime caused by defense rather than attack. Whether that trade-off holds will depend on how quickly vendors build patch pipelines that can take the volume — and on whether anyone in Washington settles on a single explanation.

Source: mlq.ai

Share this article:

More from Nathan Brooks

Nathan Brooks

Show full bio

Senior reporter covering people, launches and day-to-day practice in Semiconductors, chip manufacturing, hardware, AI compute at Chip Dispatch.

1,092 articles

Related articles

« Previous articleNext article »