Experts have warned for years that AI agents able to write computer code could seriously disrupt the digital world. Does Black October point to something worse ahead?
?
?
What is Black October?
In October 2026, a wave of serious computer hacks and cybercrime struck countries around the world, with the disruption carrying into November. The month was supposed to promote cyber-security awareness as part of a global initiative that began in the US, teaching individuals and organisations about cyber-hygiene and digital safety. The timing proved grimly fitting.
Two serious vulnerabilities in PHP - the programming language behind more than two-thirds of the world's websites - set off a global rush to protect exposed systems. Then attacks hit AI companies and developer platforms, deepening the crisis.
At month's end, thieves took £6.3 billion in digital currency from Binance, making it the largest crypto heist on record by a wide margin. Much of the blame has fallen on TraderTraitor, a North Korean hacking group. Europe's supermarket freezers defrosting was a far more basic attack, but its implications were no less troubling.
Open-weight models, particularly those coming from China, have made sophisticated hacking tools available to far more people. Their safeguards can often be stripped away - or tricked into compliance - so they assist with illegal or unethical requests. Smith's point is blunt: cybercrime no longer requires the backing of a government or a major criminal organisation. Someone may not even understand coding, websites, or what "hacking" actually involves.
Has AI made it easier to commit crime?
Yes. The overall scale of harm hackers can cause has not changed radically - at least not yet. The bigger shift is that people with little technical skill can now inflict serious damage. Defenders must deal with regulations, audits and corporate approvals, while criminals face none of those delays, James Rundle writes in The Wall Street Journal. That gives them an advantage when they exploit new tools and attack newly discovered vulnerabilities faster than any human security team can respond.
AI has not invented new attack paths, novel exploits, or new mathematics, says Adam Meyers of the US cybersecurity firm CrowdStrike. What has changed is the speed and scale: it can apply enormous computing power, then keep trying.
How should we respond?
As Cameron Peters writes in Vox, the internet age has tied systems together more tightly, which also means a failure can spread further. After Black October, individuals should get the basics right: use unique passwords, turn on multi-factor authentication wherever possible, install software updates, and treat requests for personal information with suspicion.
More broadly, companies and governments need to respond forcefully to the growing threat by creating an advantage in cyber capabilities, The Economist argues. The US government, for instance, should make it easier for companies to use cyber-defence programmes from Anthropic and OpenAI, or at least process applications from new participants more quickly. Organisations that still rely on older methods to identify and fix security weaknesses also need to start using AI, with government help where necessary.
How serious could things get?
Smith warns that losing the internet as a relatively open, inexpensive public resource - one that has driven so much economic growth - could carry radical economic and societal consequences.
Sophisticated AI cybersecurity may become a basic condition of staying in business, putting smaller organisations under severe pressure. Consumers could grow wary of online services, while financial institutions and governments retreat further into isolated networks. AI might then erode some of the internet's economic value - the same value that helped make AI possible.
That creates a huge commercial opportunity for the AI giants, with security likely to take up a larger share of their business. Everyone else may see the internet differently, though. If every company must pay Anthropic and OpenAI substantial sums just to defend against AI-powered attacks, the internet could become far less economically valuable than it has been until now. This is not dystopian fiction; it is a plausible scenario.
Leave a comment on: What happened during Black October?