Some feared an AI attack. The real reason for tech outages is more complicated
A string of technology outages across the country has the public quickly pointing fingers at artificial intelligence as fears swirl over a so-called doomsday. Researchers and public officials are suggesting AI was a likely factor — but not in the way people initially suspected. After days of uncertainty over the outages, national security officials in…
A string of technology outages across the country has the public quickly pointing fingers at artificial intelligence as fears swirl over a so-called doomsday. Researchers and public officials are suggesting AI was a likely factor — but not in the way people initially suspected.
After days of uncertainty over the outages, national security officials in the Trump administration told lawmakers on Wednesday the incidents were not due to cyberattacks but the simultaneous patching and upgrading of legacy systems.
Cybersecurity experts similarly said they so far found no evidence of coordinated malicious activity. Instead, some suggested AI’s rapid detection of new security vulnerabilities is causing various organizations to simultaneously upgrade and implement emergency patches that may not undergo full testing.
The incidents and the subsequent panic underscore a new reality of AI — one where the technology is demonstrating advanced ways to detect vulnerabilities in software and systems, while also causing temporary disruptions at the same time.
“It is easy to treat the emerging technology landscape and AI in particular as a scapegoat here, and there are certainly things that are very difficult for us to do as humans, but we’ve known for a very long time that there are challenges in the software supply chain,” Randy Rose, vice president of security operations at the Center for Internet Security (CIS), said in a media briefing Wednesday.
“There’s this kind of compounding thing that’s happening,” Rose added. “You have this accelerated AI-assisted vulnerability discovery, and you have this modernization that’s happening at the same time.”
Over the course of the past six days, outages hit various sectors, including banks, tech companies Microsoft and Apple and various government agencies. An outage at the Federal Aviation Administration led to a rare hours-long, nationwide ground stop last week, while the National Security Agency’s classified network also went down on Monday.
The simultaneous events across different industries sparked a flurry of panicked reactions online and in Washington, where lawmakers and pundits across the aisle pressed the Trump administration for answers on whether the events were related.
Amid the confusion, users online speculated the hypothetical “AI doomsday” that some of the technology’s researchers warned about had arrived, while others said they were stocking up on food and essentials in case of an infrastructure blackout.
A White House official on Wednesday said the administration’s assessment found the outages were “the result of upgrades of the systems being patched, hardware being applied,” and there is “no indication” of a cyberattack. Pointing to the companies’ stated reasons for the outages, the official said the causes are all “in the upgrading and patching realm.”
Concerns over the role of AI in the cyber world follow months of statements from Silicon Valley leaders who repeatedly warned how the technology’s capabilities could be used as both a tool to bolster U.S. infrastructure and a weapon for bad actors to exploit vulnerabilities.
After OpenAI disclosed its AI agents breached a testing environment and broke into the systems of the tech firm Hugging Face, numerous AI firms began publicly discussing similar incidents.
Vulnerability disclosures hit a record high in 2026, Rose said, but other experts said some of this is likely because AI is finding vulnerabilities that went undiscovered for years.
“Most vulnerabilities never get exploited … but we’re finding a lot of them now, and so there’s a lot of news about that kind of stuff, and that makes people worried about these kinds of things,” Nick Carroll, a cyber incident response manager at the national security firm Nightwing, told The Hill.
“It’s really just that we now have the tools that are able to actually go through this code in a way people traditionally have not had time to dig through and find all of those things,” Carroll continued.
Carroll noted the simultaneous nature of the incidents is just a matter of “unfortunate coincidences,” but he also pointed to the time of year, during which companies and teams are trying to upgrade platforms before the holidays and the year’s end.
“Everybody starts to rush, things kind of go wrong,” he said.
The rush to emergency patch newly discovered vulnerabilities could also be a factor in the recent spike in outages, experts suggested.
Before the increased use of AI, Carroll explained companies used to have a “very lengthy window” to test emergency patches, but “that’s starting to dry up for a lot of organizations,” as hackers use AI to exploit vulnerabilities more easily.
“The simple explanation is we’re finding more vulnerabilities (and faster) with AI. Fixing them is still hard and a completely different story,” Josh Bressers, the director of coordinated vulnerability disclosure at the firm VulnCheck, wrote in a LinkedIn post.
“The reality is security teams have been feeling the pressure from this onslaught all year,” Bressers wrote. “We’re well past the point of being able to ‘patch faster,’ and the uncomfortable truth is this is the new normal for the foreseeable future. Even if the volume levels off, it will be beyond what teams can handle.”
The Cybersecurity and Infrastructure Security Agency published a warning Tuesday that a critical vulnerability was discovered in common programming language Java.
The startup Calif, which used AI to discover the PHP vulnerability last month, said it used AI to discover the Java vulnerability. Java is used by nearly all of the Fortune 500 companies.
“It is the beating heart of the modern economy,” Calif wrote in a blog post Tuesday. “Banking, aviation, telecommunications, nearly every critical sector runs on Java, including national security systems.”
It is not clear whether the Java vulnerability played a role in the recent outages, though Oracle released a patch Tuesday.
Bryan Bedford, administrator of the Federal Aviation Administration, sought to quell concerns over the agency’s systems during The Hill’s “Red, Flight & Blue: Future of U.S. Aviation” event Wednesday. The agency cited a router in Arkansas that needed an upgrade, along with a Texas radar facility that caused a data flood.
“The system is absolutely safe,” Bedford told The Hill, while acknowledging, “We do have reliability challenges with our legacy systems.”
Regardless of AI’s role, cyber experts stressed companies should take away the need to bolster their security from the design phase of a product.
“The security best practices still stand,” Phyllis Lee, CIS’s vice president of security best practices, said. “It makes it even more important to implement a cybersecurity framework.”
Topics in this story
Gathered from external sources. Rights to this text belong to whoever originally published it.