The UK is losing £1.3 billion a year to scams, which works out to nearly eight fraud cases per minute. But here’s where it gets interesting: the criminals running these operations have discovered that hiring an AI to do the heavy lifting is far more efficient than the old-fashioned approach of, say, actually learning to write convincingly.
Imagine a scammer’s dilemma. You need to send personalized phishing emails to thousands of people, each one tailored enough to slip past their skepticism but generic enough that you can automate it. Enter large language models—the same technology that powers your helpful chatbot, now weaponized by someone in a basement somewhere, probably with better grammar than the original con artists.
The really amusing part is that we’ve created a world where AI can now impersonate a financial advisor well enough to convince someone to wire money, while the actual financial advisors are still trying to figure out how to explain cryptocurrency to their parents. One scammer’s AI is out there generating bespoke investment pitches; another is probably applying for a business loan to scale operations, using—you guessed it—an AI-written business plan that the bank’s algorithm approves because it looks statistically legitimate.
So the next time someone tells you AI will replace humans, remember: it already has. Not in the jobs we expected, but in the ones that matter most—making us all poorer, one perfectly worded message at a time. The real scandal isn’t that criminals are using AI. It’s that we built the tools, connected everyone to the internet, and then acted shocked when someone decided to use both at once.