Logo

Jooish

HomeSitesGroupsStatusPodcasts
Sign InSign Up
Matzav

AI SECURITY BACKFIRE: Frantic Efforts to Fix Cyber Vulnerabilities Blamed for Massive Outages at NSA, Microsoft, JPMorgan and Other Major Institutions

Nov 19, 2026·7 min read

A wave of major technology disruptions that crippled banking services, interfered with commercial aviation and temporarily disabled sensitive government communications this week appears to have been caused by efforts to repair security vulnerabilities rather than a coordinated cyberattack, according to senior federal officials and congressional sources.

Director of National Intelligence Jay Clayton told lawmakers Wednesday that investigators had found no evidence connecting the widespread failures to a foreign government or organized hacking operation. Instead, the disruptions at major corporations and federal agencies were linked to technical modifications and maintenance work, amid growing concerns that artificial intelligence is exposing weaknesses in computer networks faster than organizations can repair them.

“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 the House Homeland Security Committee.

“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,” he added.

The outages affected a remarkable range of institutions, including Microsoft, JPMorgan Chase, United Airlines, Apple Pay and the National Security Agency. Some customers were temporarily unable to access financial accounts, while aviation disruptions resulted in a nationwide halt to commercial flights. Even a highly classified NSA communications system experienced problems.

According to a report published Tuesday by the Washington Sun, senior officials from the NSA and major Wall Street institutions privately informed congressional representatives that the disruptions were connected to defensive cybersecurity work, rather than successful attacks by criminals. The report cited three individuals familiar with those discussions.

At the center of the problem is the increasingly widespread use of sophisticated artificial intelligence systems to identify vulnerabilities in computer networks. Companies and government agencies have begun deploying advanced AI models to examine their infrastructure for security flaws that could otherwise be exploited by hackers.

Although the technology allows organizations to identify weaknesses far more quickly than traditional security testing, it has also created enormous pressure to implement repairs immediately, particularly because criminals can use similar AI capabilities to discover the same vulnerabilities.

“These models are identifying massive vulnerabilities, and that increases the urgency to patch right away,” a congressional official familiar with the situation said.

“They’re trying to push through a lot of patches all at once — more than the system was designed for.”

The resulting flood of software updates and technical changes can overwhelm aging computer infrastructure, particularly when numerous modifications are introduced simultaneously without sufficient testing.

Homeland Security Secretary Markwayne Mullin raised similar concerns Tuesday during an appearance on CNBC, pointing to the difficulty of integrating advanced technology into computer systems that were built decades ago.

“They’re pushing new technology on top of really old computer systems, and it just seems to overwhelm the system,” Mullin said.

The rush to strengthen cybersecurity has intensified as federal authorities warn that artificial intelligence is also making sophisticated hacking techniques available to criminals with relatively little technical expertise.

Last week, the FBI warned that China-linked hackers were using an AI-powered security-testing framework known as ARTEX to identify vulnerabilities in banking and financial networks around the world. The technology allows users to direct automated security probes through ordinary language commands, dramatically reducing the expertise traditionally required to search for weaknesses.

“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 company DirectDefense.

“AI that can do natural language processing, and there are some that are small enough to run on a phone,” he told the New York Post.

Walcutt explained that similar technology is being used legitimately by cybersecurity professionals to identify weaknesses before criminals exploit them.

“They can turn some very powerful AI tools loose and let them. There’s things like automated penetration testing … it allows them with very little effort, basically to go hunting,” he said.

The disruption at JPMorgan Chase on Monday was attributed to software bugs, while the NSA described problems affecting its classified communications network as “scheduled routine maintenance.” Microsoft separately reported that a technical defect caused its widespread Outlook service interruption.

Although officials believe the rapid pace of security updates contributed to some of the failures, they have not established that every outage, including the Microsoft disruption and the Federal Aviation Administration’s nationwide ground stop, resulted directly from AI-driven patching.

The scale of the problem has nevertheless prompted increased scrutiny from federal financial regulators. Following the JPMorgan outage, the Treasury Department and Office of the Comptroller of the Currency contacted banks to determine whether the disruption had been contained and whether customers had been adequately informed. The OCC is also preparing a more detailed examination of the incident.

The accelerating pace of vulnerability discovery is placing additional pressure on software companies. Microsoft reportedly addressed more than 1,200 security flaws this month, nearly 20 times the number recorded during the corresponding period last year.

Cybersecurity consultant Marty Edwards, a former senior Homeland Security cybersecurity official, warned that organizations risk creating additional problems when they introduce new technology without sufficient testing and safeguards.

“We’re rushing some of the deployments of this new technology to the point where we’re bypassing some of those engineering controls,” Edwards told the Post.

Edwards also pointed to recent cyberattacks against water systems in at least a dozen states as evidence of the dangers posed by poorly secured infrastructure, particularly when sensitive industrial systems are connected directly to the internet.

“Some of those critical infrastructure systems like that are not just sitting on the internet. Now, some are. … Those systems are natively accessible from the internet, which they absolutely should not be,” he said.

“The FAA thing was a perfect example of that, because that was trying to put something new on something old,” Edwards added.

The issue has also attracted attention on Capitol Hill, where lawmakers are increasingly concerned about the broader security implications of rapidly advancing AI systems.

During Wednesday’s hearing, Rep. Matt Van Epps of Tennessee pointed to the extraordinary capabilities of newer artificial intelligence models, including their potential to assist individuals attempting to develop dangerous biological weapons.

“Leading American AI developers have activated heightened safeguards on their most capable models because they cannot rule out that those models could meaningfully help a novice build a biological weapon,” Van Epps said.

Meanwhile, Rep. Jim Himes of Connecticut, the ranking Democrat on the House Intelligence Committee, said he had reviewed the recent disruptions and found no indication that they resulted from hostile cyber activity. He cautioned, however, that increasingly powerful AI tools are enabling both attackers and defenders to discover vulnerabilities at unprecedented speeds, making rapid security updates increasingly necessary even when those updates carry operational risks.

Homeland Security Secretary Mullin told lawmakers that he was approximately 98% confident that foreign adversaries were not responsible for the recent outages, while FBI Deputy Director Christopher Raia said investigators had found no evidence of malicious cyber activity connected to the incidents.

The NSA’s classified network problems were reportedly still being addressed Wednesday, with some personnel receiving instructions to stagger their arrival times as technicians worked to restore normal operations.

{Matzav.com}

View original on Matzav
HomeSitesGroupsStatusPodcastsSign In