Widespread tech outages that wreaked havoc across the US this week were triggered by frantic attempts to patch security holes discovered by powerful AI models, federal officials revealed Wednesday.
Director of National Intelligence Jay Clayton said at a hearing of the House Homeland Security Committed that the high profile problems at Microsoft, JPMorgan Chase, United Airlines, Apple Pay and the National Security Agency were due to “maintenance and technical changes,” not 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 said.
“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.
NSA and Wall Street bigwigs also privately told Congress that the problems weren’t due to cybercriminals, but rather, efforts to thwart them, the Washington Sun reported Tuesday, citing three sources familiar with the conversations.
Major corporations and government agencies have increasingly turned the world’s most sophisticated AI models loose on their own computer networks, hunting for weaknesses that cybercriminals could exploit.
“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.
NSA and Wall Street bigwigs also privately told Congress that the problems weren’t due to cybercriminals, but rather, efforts to thwart them, the Washington Sun reported Tuesday, citing three sources familiar with the conversations.
Companies and the government are trying to stay ahead of terrifying new AI-powered hacking tools that are increasingly popping up all over the world.
The FBI warned last week that China-linked hackers are exploiting an AI testing model infiltrate banks and finance businesses around the world.
The China-made ARTEX framework can probe systems for vulnerabilities at lightning speed — allowing common crooks to simply tell it what to do, and it goes to work — with no high-level hacking skill required.
“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 cyber security firm DirectDefense, which uses similar tools to help clients test their security.
“AI that can do natural language processing, and there are some that are small enough to run on a phone,” he told The Post.
Major corporations and government agencies have increasingly turned the world’s most sophisticated AI models loose on their own computer networks, hunting for weaknesses that cybercriminals could exploit.
But the sheer volume of vulnerabilities being uncovered — and the terrifying prospect that bad guys could discover them just as quickly — has prompted a mad dash to shore up cyber defense.
“These models are identifying massive vulnerabilities, and that increases the urgency to patch right away,” a congressional official briefed on the matter told the outlet.
“They’re trying to push through a lot of patches all at once — more than the system was designed for.”
The digital mayhem of recent days has left Americans temporarily unable to access their bank accounts, forced a nationwide ground stop of commercial flights and even knocked out a highly classified NSA communications network.
Follow the Post’s coverage of the cyber outages gripping major companies:
- Outages, cyber attacks live updates: Markwayne Mullin learned of Nvidia Chinese spy scandal from the news, DHS boss tells Congress
- Trump’s top intel official reveals common cause behind outages at NSA, Microsoft, JPMorgan — and why AI is involved
- US senators warn White House that AI-enabled cyber threats are outpacing defenses
- China-linked hackers using AI to infiltrate banks, FBI warns
JPMorgan Chase blamed software bugs for its service disruptions Monday, while the notoriously secretive NSA chalked up its network woes to “scheduled routine maintenance.”
Homeland Security Secretary Markwayne Mullin pointed the finger at America’s aging technological infrastructure during an interview with CNBC Tuesday.
“They’re pushing new technology on top of really old computer systems, and it just seems to overwhelm the system,” Mullin said.
Microsoft also suffered a major Outlook outage that it attributed to a technical defect. It remains unclear whether that disruption or the Federal Aviation Administration’s meltdown was specifically linked to the AI-driven patching frenzy.
But the mad dash to patch and update is meant to stop more than just thieves and scammers, Representative Matt Van Epps, a Republican from Tennessee, said at the Wednesday hearing.
“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,” Epps argued.
And the worst may be yet to come.
As AI gets smarter, it will likely uncover even more security holes lurking in the decades-old technology that keeps America running — leaving the institutions responsible for everything from Americans’ money to their safety scrambling to fix them, experts said.
“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,” Christopher Walcutt, chief information security officer at cyber security firm DirectDefense, told The Post.
“We’re rushing some of the deployments of this new technology to the point where we’re bypassing some of those engineering controls,” said cyber security consultant and former Homeland Security cybersecurity director Marty Edwards told The Post.
Those deployments may be half-baked security updates — or simply rushed efforts to automate and transition to cloud-based systems.
Case in point: a wave of cyber attacks this summer that targeted water systems in at least a dozen states.
“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,” Edwards said.
“The FAA thing was a perfect example of that, because that was trying to put something new on something old,” he said.




No comments:
Post a Comment