What happened in Black October?
Experts have long warned that AI agents capable of writing computer code would cause serious disruption in the digital realm. Is Black October a sign of worse to come?
What is Black October?
October 2026 – “Black October” – saw a spate of serious computer hacking incidents and cybercrime across the globe – with the chaos continuing into November. Ironically, October was meant to be “cyber-security awareness month” – part of a global initiative, begun in the US, to educate individuals and organisations on cyber-hygiene and digital safety. It turned out the timing was all too horribly appropriate.
Most worryingly, two critical vulnerabilities in PHP, the programming language that runs more than two-thirds of the world's websites, triggered a worldwide scramble to secure affected systems. Numerous attacks on AI firms and developer platforms compounded the sense of crisis.
Then, at the end of the month, $6.3 billion in digital currency was stolen from Binance, the cryptocurrency exchange, in by far the biggest crypto heist to date. The North Korean hacking group TraderTraitor has been widely blamed. By comparison, supermarket freezers defrosting across Europe was a low-tech kind of attack – but no less ominous in terms of the implications.
Start your trial of MoneyWeek today
Get unparalleled financial insight, analysis and expert opinion you can profit from.
Sign up for MoneyWeek’s free twice-daily newsletter.
Join more than 165,000 subscribers and keep yourself informed with latest financial news, insights and expert analysis to help you understand what really matters when it comes to your finances.
Join more than 165,000 subscribers and keep yourself informed with latest financial news, insights and expert analysis to help you understand what really matters when it comes to your finances.
Were only businesses affected by Black October?
No, governments and states have been attacked, too. In Ukraine, enemy hackers accessed national emergency apps used to warn the population of imminent danger, issuing fake warnings of nuclear attacks. In response, similar hacking attacks were soon seen in Russia.
In Denmark, cybercriminals stole data from the official population registry. In Japan, a series of attacks prompted the government to declare a “cyber-security state of emergency”. In the US, the Department of Homeland Security and the FBI were both the targets of data theft.
In the UK, the National Cyber Security Centre warned last week that hackers are highly likely to use AI over the next two years to launch cyberattacks on critical infrastructure on a faster and far larger in scale than in the past.
Has AI played a role in the cyberattacks?
AI has made hacking far easier to do and far more accessible to bad actors. For months, says Noah Smith on Substack, experts had been warning that the arrival of AI agents that are extremely good at coding was going to cause serious disruptions to the digital realm sometime this autumn. And now it's happening.
The internet – the digital infrastructure on which modern life and commerce depend – is inevitably full of the coding flaws, mistakes and vulnerabilities produced by “sloppy and fallible” humans. Now that AI agents are very good at coding, they are capable of exploiting those flaws.
And the open-weight models, especially from China, have “democratised access to high-quality hacking”. These models can be fairly easily “abliterated” – meaning their guardrails removed or deceived so that they will comply with illegal or unethical demands. In other words, says Smith, you “don't need the resources of a nation-state or even a big organised crime ring to commit cybercrimes now. You don't even have to know how to code, or how websites work, or what ‘hacking' even really means.”
Has AI made it easier to commit crime?
Yes. So far there hasn't been a radical shift in terms of the scale of the damage hackers can inflict. It's more that even unskilled hackers can now do serious damage. Bad actors “don't face the same restrictions that impede defenders, including regulations, audits and corporate approvals”, says James Rundle in The Wall Street Journal. They've got an in-built advantage when it comes to exploiting the newest tools and targeting newly discovered vulnerabilities at a speed no human security team can match.
“AI hasn't come up with new attack paths. It hasn't come up with novel exploits. It hasn't come up with new maths,” says Adam Meyers of US cyber-security firm CrowdStrike. The difference is simply its ability to deploy vast computing power at speed and keep trying.
How should we respond?
It's long been well known that the greater interconnectedness of the internet age creates greater systemic risk, says Cameron Peters in Vox. The best individual response to Black October is that we all start doing the basics well – changing passwords and not reusing them, using multi-factor authentication wherever possible, keeping software updated and remaining suspicious of requests for personal information, and so on.
More broadly, companies and governments must act decisively against the emerging threats by “generating an imbalance in cyber-weaponry”, says The Economist. For example, the US government should take a more permissive attitude to corporations wanting access to the cyber-defence programmes offered by the likes of Anthropic and OpenAI, or at least approve new joiners faster. And organisations that are not yet using AI to find and close security holes must catch up – and get government support if needed.
How serious could things get?
If we can not rely on the internet as a relatively open, inexpensive public resource that has driven so much economic growth, then there are radical economic and societal consequences, says Smith.
Every business will need sophisticated AI cyber-security just to stay operational, crushing smaller organisations; consumers will become reluctant to trust online services; and financial institutions and governments will increasingly retreat into isolated networks. In other words, there's a risk that AI could destroy some of the economic value created by the internet that made AI possible in the first place.
That's a gigantic commercial opportunity for the AI giants – security will become an increasingly important chunk of their business. But for the rest of us, an internet in which every company has to pay substantial sums to AI companies such as Anthropic and OpenAI – simply to protect itself from AI-powered attacks – would be much less economically valuable than up until now. That's not a dystopian fiction; it's a very plausible scenario.
This article was first published in MoneyWeek's magazine. Enjoy exclusive early access to news, opinion and analysis from our team of financial experts with a MoneyWeek subscription.
Join more than 165,000 subscribers and keep yourself informed with the latest financial news, insights and expert analysis to help you understand what really matters when it comes to your finances.