U.S. Sen. Mark Warner told Americans on Tuesday that the week of failures that rattled their banks, airports and email was not an attack. In a Nov. 17 statement, the Virginia Democrat, vice chairman of the Senate Intelligence Committee and co-chair of the Senate Cybersecurity Caucus, said there was no indication the incidents resulted from a cyberattack, foreign interference or a coordinated effort. He pointed to something harder to fight: the speed of artificial intelligence itself.
Warner’s argument runs like this. AI can now find vulnerabilities across enormous numbers of systems at unprecedented speed, pressuring companies and agencies to patch them immediately, “often all at once, putting tremendous strain on systems that weren’t built to handle that pace of change.” He added, “While we still need to get to the bottom of these incidents, the evidence so far points toward legitimate efforts to secure our systems, not malicious activity,” he said. He called the outages a wake-up call, warning that “AI is moving faster than our ability to manage its consequences.”
Warner, who has spent years warning about AI’s security risks, framed his assessment as preliminary and said the incidents still need to be fully investigated. The initial explanations from the organizations involved point to a range of technical causes. Microsoft cited “a technical issue not related to cyber or AI;” while JPMorgan Chase claimed a software bug; the FAA stated new routers; and the National Security Agency blamed scheduled maintenance. Final incident reports should show whether rushed security fixes played any role.
The Trump administration’s message has been blunter. Director of National Intelligence Jay Clayton told CNBC on Tuesday, “This is not a cyberattack.” Secretary of State Marco Rubio, speaking in Manila, said there was “zero evidence” the incidents were related and that each had its own technical cause.
The most consequential failure came Friday, Nov. 13. at about 3:30 p.m., when the FAA ordered a nationwide ground stop after its legacy telecommunications network was flooded with data, destabilizing radar and radio feeds at 17 of its 22 en route control centers. Transportation Secretary Sean Duffy blamed a new router installed in El Dorado, Ark., by contractor L3Harris during a copper-to-fiber upgrade. The stop lifted around 8:15 p.m., but cancellations ran into the thousands through the weekend. At four Virginia airports, including Reagan National and Dulles, 532 flights were affected, the Virginia Mercury reported.
The FAA says the root cause is still under investigation; it has paused installation of the new Juniper routers and disabled the 68 already in service. FBI spokesman Ben Williamson denied a CNN report that agents had seized the equipment, saying the bureau is helping examine hardware that remains with the provider. As described by Duffy, the FBI’s job is to make sure the system was not hacked, which is precautionary technical help, not a criminal case. No agency has released forensic findings.
The ground stop hit a region already uneasy about technology in the control room. In September, Rep. Don Beyer, whose district includes Reagan National, objected to an AI-assisted traffic management tool the FAA rolled out at Washington-area airports, warning against using his constituents “as guinea pigs for an unproven AI air traffic control system,” as the News-Press has reported. Nothing in the public record ties that tool to the Nov. 13 failure. The two share only a theme: new technology pushed into old systems faster than they can absorb it.
On Monday, Nov. 16, JPMorgan Chase customers spent much of the day unable to log in or move money, and the trouble spread to some branches and to outside apps linked to Chase accounts. The bank blamed a software bug in its on-premises systems, and a source told CNN it was neither cloud-related nor a cyberattack. Customers at Wells Fargo, Capital One, Bank of America, U.S. Bank and Zelle also reported trouble, but whether those were separate failures or ripple effects from Chase has not been established.
Microsoft’s outage, which began Nov. 12, knocked out Outlook, Teams and other Microsoft 365 services for more than a day. Its preliminary report blamed a months-old server bug exposed by routine network changes. Apple’s iCloud mail, photos and sign-in services failed for hours on the evening of Nov. 10, and Apple has not disclosed a cause. Separate failures, but together they show how much runs through a handful of providers where one bad change reaches millions.
On Tuesday, CNN reported that NSANet, the NSA’s classified internal network, was down. The agency said it suffered a “temporary degradation” during scheduled maintenance and that mission-critical capabilities stayed operational. It has not said how long the outage lasted, and there is no public evidence that classified information was compromised. That evening, Fairfax County Fire and Rescue crews responded to CIA headquarters in McLean after reports of smoke. Crews traced it to the building’s HVAC system and found no fire or other emergency.
The most confusing signal came from the FBI. On Nov. 13, the bureau issued a FLASH advisory warning that criminals and state-aligned hackers have used ARTEX, an AI-powered penetration-testing tool, to automate attacks since August, some reaching privileged access in under three hours. When FBI Dallas reposted it Tuesday, it drew millions of views, and headquarters clarified that the advisory “is not a new announcement.” The Dallas Morning News reported the tool has not been linked to the outages.
The FBI and the officials denying an attack were not contradicting each other. One message said malicious actors are using AI to break into networks; the other said these failures were not their work. But an alert issued on the day of a nationwide ground stop, then amplified without context, left the public to connect dots officials say do not connect. “Americans deserve clear, reliable information and a government that is prepared to respond,” Warner said.
All of this unfolded while the government’s lead civilian cyber defender is far smaller than it was two years ago. Layoffs, buyouts and other departures cut the Cybersecurity and Infrastructure Security Agency from roughly 3,300 employees to about 2,200, according to House Democrats who in August asked the Government Accountability Office to determine what capabilities were lost. The administration has defended the cuts as refocusing CISA; a DHS statement in October 2025 said they were part of “getting CISA back on mission.” Warner sees it differently, saying Trump “has shown little interest in confronting these challenges” and that the CISA cuts “have weakened the very expertise we need to navigate moments like this.”
Now the administration appears to be partly reversing course. Acting Director Nick Andersen has said CISA plans to hire about 330 people, and Homeland Security Secretary Markwayne Mullin told Congress the ideal size is roughly 2,800, still below the 3,400 positions authorized before the cuts. Yet the fiscal 2027 budget proposes cutting more than $700 million from the agency. On Nov. 9, President Trump nominated Michael LoSapio, Palantir’s chief information security officer, to lead a CISA that has gone nearly 22 months without a Senate-confirmed director.
There is no evidence the staffing cuts caused any of this month’s outages, most of which hit private companies or agencies outside CISA’s control. The accountability question is different: whether the country shrank its civilian cyber capacity just as AI made defending critical systems harder, and whether rehiring a few hundred people while proposing deeper budget cuts adds up to a strategy.
“I’m going to keep digging into these incidents, pressing for answers,” Warner said, and he has a legislative agenda, though none of it is law. His “Framework for America’s AI Future,” unveiled in July, includes the Secure AI Development Act, requiring secure government testing of frontier AI models. A separate bill introduced Sept. 29 would create an AI Safety Board at the Commerce Department and require critical infrastructure operators to report AI safety and security incidents, within 72 hours when the threat is imminent. Neither has passed, and neither would have stopped a router from flooding the FAA’s network. But mandatory incident reporting goes to this week’s core problem: the public waited days to learn what broke.
For Falls Church, none of this is abstract. Northern Virginia hosts the airports, intelligence agencies, contractors and data centers at the center of this story. A Falls Church resident could have been locked out of a Chase account, stranded at Dulles and cut off from work email in the same five days. The questions for local government are concrete: what keeps 911 dispatch, payroll, utility billing and schools running when a major provider goes dark, and when were those backup plans last tested?
Five failures in one week can have five causes, and so far the evidence suggests they did. That is not entirely reassuring. A router upgrade, a latent server bug, a software release and routine maintenance each disrupted millions of people without any adversary lifting a finger. Accountability would mean public incident reports, an independent review of the FAA failure and oversight hearings. Whether America’s infrastructure is resilient enough for the technology it now depends on is the question this week raised, and no one in Washington has answered it yet.
