Outages in US cause worry, despite official attempts to downplay concerns
Consumers in the United States have dealt with a number of technology-related outages in the last week, affecting services provided by major banks and companies such as Microsoft, Workday, and Apple.
Even the US government has experienced problems, including a national aviation ground stop and an ongoing problem affecting the internal network at the country’s premier surveillance agency.
Recommended Stories
list of 3 items
- list 1 of 3US legislators push AI safety laws amid human extinction warnings
- list 2 of 3Why AI could threaten much of the world’s digital infrastructure
- list 3 of 3Microsoft hit by global outages: What happened and why?
end of list
US officials have downplayed suggestions that the incidents are related.
“Tell them to relax, everything is good, these are individual events,” US President Donald Trump told a reporter on Fox on Thursday when asked about the outages and whether the public should be preparing for emergencies.
The White House fired the US cybersecurity chief on Thursday and appointed a replacement, days after the disruptions.
US Director of National Intelligence Jay Clayton said in media interviews on Wednesday that the outages likely stemmed from organisations using artificial intelligence (AI) to fix security flaws and carry out maintenance updates.
The tech outages come after months of public alarm over the threat posed by AI, as well as messaging by governments in the US and other countries about other events that has inadvertently added to that concern.
In the UK, Sweden, and Japan, for example, authorities have run campaigns in recent months to remind the public to store essentials at home such as torches, water, and medicine, in case they experience stronger-than-usual storms this winter due to the El Nino phenomenon. Last month, the US Department of Defense announced it was rolling out a programme to ensure all military bases could operate for 90 days without relying on the civilian power grid.
Advertisement
Heightened risk from AI
“It’s been an unusual stretch. People have had problems with several services they rely on in their daily lives, and we’re still learning what happened in some of these cases,” said Danny Atherton, a lecturer at Georgetown University and editor of the AI Incident Database, which tracks global outages and attacks reported to have involved the use of AI.
“I worry most about essential services,” Atherton said. “If a shared system goes down, can a hospital still take care of patients? … I think governments and AI providers should expect serious outages to continue.”
More and more companies are using AI to test their systems and finding vulnerabilities that could potentially be exploited by malicious actors in unexpected ways.
Last month, the FBI said cybercriminals linked to the North Korean government were able to steal $6bn in cryptocurrency from the Binance crypto exchange. The hackers reportedly used AI in several parts of the complex operation, posing as recruiters from another company to conduct video interviews with engineers, planting malicious software on their computers, and eventually gaining access to crypto wallets holding the funds.
Peter Slattery, a research scientist at MIT’s FutureTech, said the Binance heist is one example of how AI might have played a role in the series of outages in recent weeks, but may not yet have been noticed. Malicious actors could have used AI to conduct research, or create deepfakes that were convincing enough to win the trust of key employees at target companies. Software may have been updated to accommodate AI tools, or it might have recommended changes to how a company operates that turned out to be problematic.
“We are moving into an era where AI is increasingly useful, both for hacking and for a variety of other tools,” Slattery told Al Jazeera. “It’s quite likely, but not guaranteed, that things will get worse. We will have more attacks from AI on infrastructure, and there will be some significant attacks and harmful events over the next year or two.”
Slattery points out that companies are already seeking to insulate themselves from liability if a potential AI-powered attack happens. In the last few months, for instance, several major banks, including JP Morgan, Citibank, and Bank of America, have notified customers that if they use third-party tools that use AI to access their accounts, the banks will not be liable for any losses if something goes wrong.
Underestimating threats
The speed with which Trump administration officials have brushed off questions about the threat posed by recent outages has also reduced confidence among experts, like Deanne Criswell, who served as the head of the Federal Emergency Management Agency (FEMA) until 2025.
Advertisement
“We can’t just take the first answers as the absolute answers, because sometimes these vulnerabilities can be buried deep within the system, and we have to do our due diligence to understand the cause,” Criswell told Al Jazeera.
She said that incidents like the airline ground stop last week should not be treated in the same way as natural disasters. After a hurricane, FEMA officials can fairly quickly assess the damage, and estimate how long it would take to restore critical services, whether that is days, or weeks, or even months later.
“But in a cyber or tech-related incident, I don’t know that, I’m not an expert in that, and I have to bring in those outside partners … to help us better understand the cause,” Criswell explained.
FEMA relied on the Cybersecurity & Infrastructure Security Agency (CISA) to do that, Criswell said, but it is unclear what the impact of budget and staff cuts has been at the agency under Trump.
The agency is tasked with maintaining contact with the myriad of private companies that control everything from water to power and telecommunications in the US. More than a third of its staff were lost to cuts by the Department of Government Efficiency (DOGE) in 2025, and in the administration’s proposed budget for 2027, CISA would face $700m in cuts.
The measures demonstrate “a dangerous underestimation of the threats facing our nation from adversaries and criminals who seek to destabilise our national security, economy, public health, and safety”, Democratic Senator Mark Warner of the state of Virginia wrote to the agency in June.
Emergency preparedness becoming more mainstream
Regardless of official messaging, or what might actually be to blame for the incidents, they are not going unnoticed by the US public, said Chris Ellis, a disaster preparedness expert and Director of Education and Strategic Outreach for The American Civil Defense Association.
More and more people in the US have sought to collect supplies such as food, water, and other essentials in recent years, especially after moments of crisis, like the outbreak of the Ukraine war, or the global pandemic. After each election, Ellis says, the losing side in the US has also tended to purchase emergency supplies. In 2024, more than 20 million Americans had enough supplies to last them 31 days or longer after a disaster, according to Ellis – a number that has grown in recent years and months.
Regardless of government efforts to calm public fears, Ellis says the proportion of people who are actively preparing for a disaster is reaching the point where the practice might be normalised, as it was in previous generations, or still is in some other countries.
“We are kind of at that stage where prepping is still seen as fringe, but not crazy,” he said.
Related News
Ilhan Omar plans to move out of her Washington DC home after federal raid
Lagos Art Week confronts history, identity and belonging
Exit polls following Israeli elections indicate no clear winner