Fridges across Europe roasting their contents. Microsoft emails going offline for hours. US airports grinding to a halt.
The world has seen a seemingly unprecedented spate of IT blackouts, tech failures and cyber attacks in the last few weeks.
Billions of dollars have been stolen in cyber attacks while critical vulnerabilities in widely used software forced government websites offline for hours last month.
The frequency of incidents led IT professionals to dub last month “Black October”. But if anything, the rate has only increased since then.
America’s national cyber centre warned companies on Tuesday to fix a “critical vulnerability” in Java, a coding language used across the web.
Cyber incidents – stolen data, ransoms, and leaked messages – can often feel abstract. Yet in many of the recent cases they have had visible real-world consequences.
A Google Cloud problem in October knocked banks offline for hours. Last week’s Microsoft outage meant NHS trusts resorting to pen and paper. Fridge attacks in France , Belgium and elsewhere meant millions of euros of food having to be thrown away.
Nicole Perlroth, a cyber expert, said she had received hundreds of messages “asking me if the world is, indeed, ending”.
The cascade has triggered theories that the West is under a co-ordinated and far-reaching cyber attack intended to bring down an increasingly internet-dependent society.
Conspicuously, it came in the same month that the British Government urged households to stockpile water, food and batteries.
Josh Hawley, a US Republican senator, said on Tuesday: “If this is all a coincidence, it’s like the greatest coincidence in the history of the world.”
Officials have dismissed concerns. Marco Rubio, the US secretary of state, said: “We are not under systemic attack. These events, one has nothing to do with the other, they are not interrelated.”
Jay Clayton, the US director of national intelligence, said: “We do not see the indicia of a coordinated cyber attack. And we do not see the indicia of state- activity here.”
Donald Trump, when asked them, said: “It’s fine.”
There is no evidence that the attacks, outages and vulnerabilities are the result of a single campaign. Some have been attributed to foreign states, such as the fridge attacks, which have been blamed on Iran. Others have more prosaic explanations such as a software update gone wrong.
But neither is the frequency of these problems necessarily a coincidence. A combination of AI-assisted hacking, ageing IT infrastructure and a rush to bolster computer systems has created ample opportunities for things to go wrong.
“We’re seeing a triangle of fire when it comes to cyber incidents,” says Ilkka Turunen of security company Sonatype, warning that the state of play might be the “new normal”.
Cyber experts have been warning for months that AI could lead to a wave of hacking attempts. In April, Claude maker Anthropic unveiled Mythos , which it warned would “reshape cybersecurity” due to its ability to discover bugs. At the time, security chiefs warned that within months criminals would get their hands on similar technology.
The predictions proved correct. This week, security researchers at Google said Russian-speaking hackers had pulled off the world’s first fully automated cyber attack.
“Inexperienced cybercriminals can execute attacks at the level of nation state actors nine months ago,” says Turunen.
None of the major IT problems of the last few weeks has been solely attributed to malicious AI. What has been more disruptive is a wave of bugs discovered by cyber defenders.
In October, researchers at the US security company Calif said they had used AI tools to discover a critical bug in PHP , a widely used web language used to access databases. The rush to fix the flaw, and concerns that it could be abused, sent dozens of government websites offline in addition to hundreds of other sites.
The same company this week revealed the Java flaw, which attackers could use to crash IT systems. Both flaws were found as part of coordinated work between Calif and OpenAI, the maker of ChatGPT.
Cyber defenders have been given privileged access to powerful AI hacking tools to give them a head start on cyber criminals. But this has led to a flood of newly discovered flaws and pressure to fix them, even if they might never have been exploited.
More than 11,000 critical software vulnerabilities have been discovered so far this year, according to the US National Vulnerability Database. This is up from 3,365 last year, and represents more than a quarter of the bugs ever discovered.
That has, in turn, led to a breakneck effort to fix them before they are discovered by criminals. However, updates do not always work: an error in new code or failure to integrate with a separate piece of software can crash a whole system.
Microsoft, which blamed a faulty patch installed earlier this year for last Friday’s outage, has seen the number of weekly bug fixes it is undertaking quadruple this year.
“We’re going through this patch apocalypse right now. And sometimes when you’re updating systems, we all know things can go wrong,” Dmitri Alperovitch, the co-founder of security company Crowdstrike, told NPR on Tuesday.
“At some point the dam is going to break,” Marcus Hutchins, the security researcher who defeated the 2017 WannaCry virus, warned at the weekend. He said staff tasked with maintaining software were facing an “endless torrent of bug reports” and were “just going to burn out and give up”.
Warnings of an AI hacking wave have finally provided an impetus for ageing and obsolete IT systems to be updated, but this is often a painful and disruptive process. Ollie Whitehouse, the head of the National Cyber Security Centre, warned in May that such “legacy” systems might need to be completely replaced.
Upgrades to ageing systems were behind this weekend’s US airports’ shutdown, as well as the US National Security Agency’s secure communications system.
White House officials said on Wednesday that the spate of outages was due to the “hardening and upgrading of our systems”, not a wave of cyber attacks.
This may not be totally reassuring. If this is the chaos caused by the good guys, imagine the damage that the promised wave of criminal AI attacks could do.
The full story
This article is one source in a clustered incident — the cluster page carries the summary, timeline and every other outlet covering it.
