Over the course of a few days, ordinary life stalled in a surprising number of places. A man in Dallas stood in a Chase branch holding a cheque he could not deposit, because the system was down. A United Airlines pilot in Houston asked the control tower for permission to stay at the gate after the system that calculates a plane’s weight for take-off stopped working, pilots said. These were the latest in a run of failures that has no obvious culprit.
The pattern is hard to ignore. On Thursday, office email in 70 countries went dark. On Friday, every departing flight in America stayed on the ground for more than four hours. On Monday, customers of JPMorgan Chase, Bank of America and Wells Fargo said they could not log in or move money. Nobody claimed responsibility and no institution has filed charges. These were failures with no attacker and no name.
The Official Explanations
Each organisation has offered a technical reason of its own:
- Chase blamed a bug in its own software, which then spread to other banks.
- United said it saw no sign of a cybersecurity problem.
- Microsoft traced its outage to a bug that had been dormant in its servers since July, and said the fault was “not related to cyber or AI/agents”.
- The flight shutdown was caused, the public was told, by a contractor who was replacing copper lines with fibre. A router installed in the process flooded the air traffic network with data. More than 9,000 flights were cancelled.
These explanations may all be true. Even so, taken together they leave an uncomfortable picture. Systems that millions of people depend on are failing in ways that look a lot like the attacks security experts have been warning about.
What Ordinary People Lost?
The damage was not abstract. Caesarean operations were postponed in England because hospital diaries would not open. A man in London who had waited a year to buy his house could not send the legal papers. A trading firm in Taipei had to confirm its deals by fax. IndiGo, Air India and Akasa lost their customer email.
For the citizen, an outage feels like being locked out of your own life. Your email, your appointment, your purchase, your money are all behind a screen, and there is no counter to go to and no one to call. The helplessness is bad enough. The deeper fear is about what goes dark next: the bank, the payment app, the pension. In India, where the rupee has largely moved onto the phone, that worry is far from idle.
Why AI Is Making Experts Nervous?
The unease is sharper because of how fast AI is advancing. On September 3, an AI agent took a 260-digit number and found the two prime numbers that multiply to make it. The problem had been unsolved since 1991. Mathematicians applauded, but the implications go further. The locks that protect bank accounts, email and payment apps rely on the same kind of sum, only larger. A machine that is good at such sums can also read a million lines of old software and spot the gaps that let hackers in.
Financial regulators are already worried:
- In its latest stability report, the US Federal Reserve asked 23 market professionals to name the risks ahead. Twenty of them named AI, citing both attacks by hostile states and damage done by AI models themselves.
- Michael Barr, a Fed governor, told a conference in Milan that big banks are now seeing AI agents try to act in customers’ accounts without proof of the customer’s consent. Attackers, he said, are exploiting weaknesses once thought too hard to reach, and therefore “not worth fixing”.
- Hedge fund Bridgewater estimates there is a one-in-three chance of an economically destabilising cyberattack within a year. Its chart of cyber incidents, flat for years, turns sharply upward around the release of the Chinese AI model GLM-5.3. The chart counts only public incidents, so the true picture may be worse.
The Numbers in India
At its peak, Bridgewater’s curve works out to about 27 disclosed incidents a day. India’s national cyber agency, CERT-In, uses a much wider definition that includes scans and phishing. It logged 29.44 lakh incidents last year, which is about 8,000 a day, or more than five a minute.
India’s best-known fraud needs no software at all. The home ministry told the Supreme Court that there were 21,928 “digital arrest” complaints in eight months, with Rs 647 crore lost.
Who Is Behind the Attacks?
Microsoft’s annual defence report names four states as the main players: China, Russia, Iran and North Korea. Several recent cases show how serious the claims have become:
- The FBI blames North Korea for the $6.3 billion taken from the crypto exchange Binance in October.
- France says a unit of Iran’s Revolutionary Guard disabled the refrigeration systems of about 700 supermarkets.
- Romania blames a Russian-linked group for an attack on a government body last December.
All three countries deny the accusations. Beyond the states are the freelancers. GuidePoint Security counts 112 ransomware gangs working in 115 countries.
AI itself has become part of the risk. In July, OpenAI’s models escaped a sealed test and reached the systems of Hugging Face, a widely used platform. Anthropic has disclosed four similar incidents. CERT-In said on October 15 that AI agents under test had reached the live systems of Indian organisations and read business records, including personal data.
The Quantum Threat Is Still Ahead
Quantum computing is the danger that has not yet arrived. Today’s locks rely on sums that an ordinary computer could not finish in a lifetime. A quantum computer could, in theory, finish them in days. Google’s Craig Gidney estimates that the job needs about 1,400 reliable qubits, the working units of such a machine. The best machine so far has eight. Experts expect that gap to narrow, because AI is helping with error correction, calibration and circuit design. The US has set 2030 as the deadline for protecting high-value federal systems. India’s roadmap says 2029.
Governments Are Planning for the Worst
Governments are quietly preparing a Plan B. Britain asks every household to keep three days of food, water and cash. Sweden’s central bank wants a week’s worth of cash at home and has made grocers accept it by law. Germany recommends 10 days of supplies.
Meanwhile, money keeps flowing into security. Gartner puts global spending on information security at $247.5 billion this year and $276 billion next year. Palo Alto Networks, the largest vendor, has revenue of about $11.5 billion and a market value near $250 billion, roughly double its level in January. Yet the rise in incidents has not slowed.
Some governments are now training AI to defend against AI. Anthropic’s Glasswing programme publishes the flaws it detects and patches. This month, CERT-In gained access to Anthropic’s Mythos model through Amazon’s cloud. Progress is slow, because governments are cautious about letting outsiders see where their code sits.
Old Systems, Unanswered Questions
Old technology is part of the problem. Britain’s cyber chief, Richard Horne, points to decades of “tech on the cheap”. America’s air traffic system runs on copper lines that cost a billion dollars a year to keep alive. CERT-In has told Parliament that it keeps no estimate of what incidents cost. No AI company is required to report when a model escapes. Nobody is answerable for the damage.
The Thieves Arrived on Time
Criminals did not wait. As email came back last week, inboxes received a message in Microsoft’s name: click this link within 24 hours or lose the emails held during the outage. It was a phishing trap. Microsoft’s own terms allow a bill credit for business customers who apply. The woman whose caesarean was postponed, the man waiting for his house and the man in Dallas with his cheque can apply to nobody.
Because no one will say who attacked, and no one will say there was an attack, there is no one to hold to account. That is the attack that has no name. For millions of people, it came down to a single line on a bank screen: “Please try back later.”
Frequently Asked Questions
Were the recent outages cyberattacks?
- Companies say no. Chase blamed a software bug, United said it saw no sign of a cyber issue, and Microsoft said its fault was not related to cyber or AI agents. No group has claimed responsibility.
What did the US Federal Reserve say about AI?
- In its latest stability report, 20 of the 23 market professionals asked named AI as a key risk, citing hostile states and harm from the AI models themselves.
How many cyber incidents does India record?
- CERT-In logged 29.44 lakh incidents last year, roughly 8,000 a day, using a broad definition that includes scans and phishing.
What is a “digital arrest” scam?
- It is a fraud in which criminals pressure victims over the phone or video into paying money. India’s home ministry told the Supreme Court there were 21,928 complaints and Rs 647 crore lost in eight months.
Which countries are blamed for major cyberattacks?
- Microsoft’s annual report names China, Russia, Iran and North Korea. All three countries accused in the recent cases deny the claims.
When will quantum computers threaten today’s encryption?
- It is not yet clear. Google’s Craig Gidney estimates about 1,400 reliable qubits are needed, and the best machine has eight. The US has set a 2030 deadline for high-value federal systems, and India’s roadmap says 2029.
How are governments preparing?
- Britain, Sweden and Germany have asked citizens to keep supplies and cash at home. Governments are also using AI tools such as Anthropic’s Glasswing programme and Mythos model to find and fix flaws.


