Every time you search for a pair of shoes and then see those exact shoes following you across every website for the next week, you’re looking at AI in action. It’s efficient. It’s also a little unsettling when you stop and think about it.
AI has become the engine behind most digital advertising today. Algorithms decide what you see, when you see it, and how often. Brands no longer just buy ad space they buy predictions about human behaviour. And those predictions are getting frighteningly accurate.
That’s the part most people don’t talk about when they celebrate how “personalized” ads have become.
Here’s what AI-powered advertising actually does: it collects data about what you click, how long you hover over something, what you buy, where you go, what time you’re usually on your phone and uses all of that to build a profile of you. Not a rough sketch. A detailed one.
That profile gets used to show you ads at moments when you’re most likely to act on them. Sounds useful on the surface. But there’s a line between helpful and manipulative, and it’s not always clear where AI draws it.
Researchers at Princeton and the University of Washington have documented cases where ad algorithms ended up showing high-paying job ads mostly to men, and predatory loan ads more frequently to people in low-income zip codes. The algorithm wasn’t “told” to discriminate it learned to, because that’s what the historical data reflected. The bias was baked in before anyone noticed.
This is one of the more serious problems with AI in advertising. The system optimizes for conversions, not fairness.
Most people agree to cookies and data collection without reading what they’re agreeing to. That’s not really a secret. Privacy policies are written to be long, not clear.
What’s changed with AI is the scale and sophistication of what happens after that click. Your browsing habits, location data, social media activity, and purchase history can be stitched together in ways that most people never imagined when they ticked “I agree.” The ad you see isn’t random it’s the output of a system that has spent a lot of time figuring out what you want before you’ve figured it out yourself.
There’s something uncomfortable about that even if you can’t quite articulate why. It’s not that you mind seeing relevant ads. It’s that you didn’t sign up to be profiled this deeply. The consent you gave and the consent that was actually needed are two different things.
The European Union’s General Data Protection Regulation (GDPR) was an attempt to close that gap. It’s had some impact. But enforcement is uneven and companies are creative about working around the spirit of the law while technically following the letter of it.
There’s another layer to this that’s still developing. AI can now generate faces, voices, and personalities that don’t exist. Some companies are already using synthetic AI-generated influencers to promote products. These digital personas have millions of followers on Instagram. Their audiences often don’t know or don’t particularly care that they’re not real.
This raises a question worth taking seriously: if a brand can manufacture a “person” whose entire purpose is to sell things, what happens to trust? What happens to the idea that a recommendation means something?
Advertising has always involved some degree of constructed reality. But there’s a difference between a well-lit product photo and an entirely fabricated human being. Where that line sits legally and ethically is still being worked out.
The technology is outpacing the regulation. That’s not new in the history of media TV advertising went through similar growing pains. But the speed of AI development means the gap is wider this time.
A few things seem clear. Advertisers using AI should be required to disclose when targeting is based on sensitive inferences emotional state, financial stress, health concerns. Algorithmic systems should be audited for discriminatory outcomes, not just technical accuracy. And synthetic media used in advertising should be clearly labelled.
None of this is simple to enforce. But the conversation needs to happen now, not after the problems become impossible to ignore.
AI in digital advertising isn’t inherently bad. It can genuinely make advertising more relevant and less wasteful. But relevance and manipulation can look identical from the outside, and right now the industry is doing very little to help people tell them apart.
That’s the ethical concern. Not the technology itself what it’s being used for, and who decides.