AI-powered behavioural profiling¶
It’s not just profiling. It’s pattern gardening — and you’re the tomato.
Once upon a time, advertisers had to guess who you were based on your choice of newspaper or biscuit. Now, thanks to AI and more data than sense, they know more about you than your gran — and unlike your gran, they’re monetising it.
Modern data harvesters — think advertisers, brokers, and bored machine learning interns — don’t just collect the obvious stuff like what you search for or click. They correlate. That means stitching together your online habits (scrolling, shopping, streaming, doom-clicking) with your offline life (credit card swipes, location data, gym check-ins, what you had for lunch, when you last cried in a Tesco).
Out of this tangled hedge of data, they build “super-profiles”: unnervingly detailed portraits of your preferences, vulnerabilities, routines, and maybe your blood pressure too. All without you ever typing a word.
Example? Retailers have famously predicted pregnancies based on purchasing patterns before the customers themselves made it public. A teenager’s dad found out she was expecting because she started getting baby product coupons in the post. That’s not marketing — that’s surveillance with a loyalty card.
These profiling techniques are often powered by black-box AI models that aren’t just hard to interpret — they’re deliberately opaque. The AI doesn’t “understand” you in a meaningful way, but it does make high-stakes predictions about what you might do next — and whether you’re profitable, insurable, or risky.
More examples of what this might lead to:
Your dating app preferences subtly adjust based on shopping history or mobility data.
Health insurers infer your lifestyle based on your Deliveroo orders and Apple Watch data.
Mortgage lenders flag you as a bad risk because you frequent a particular pub too often.
And you can’t opt out. You often don’t even know it’s happening.