ChatGPT and ads, welcome in the hyerpersonnal target aera

Elisa Trousson
24 hours ago
5 min read

Images des publicités de Claude lors du Supoer Bowl 2026 contre CHATGPT

While everyone was watching the Super Bowl in early February 2026, another battle was playing out on screens. Anthropic, the company behind Claude, launched ads mocking ChatGPT and its decision to integrate advertising. A fitness coach pushing you to buy insoles to appear taller, a teacher slipping jewelry promotions into grading feedback… The message was clear: ads are coming to AI, and it’s going to be weird.

But behind this clash between two AI giants lies a much deeper issue than simple advertising. We’re talking about a radical shift in how brands will target you. Not just “targeted” like today, but hyper-personalized—and that distinction is huge.

What we experience today: category-based targeting

Currently, when you scroll on Instagram or YouTube, you see ads. Nothing new. These ads are supposed to match you because algorithms have placed you into categories. You’re a woman in your thirties? Boom—ads for Clearblue pregnancy tests. You searched for a couch on Google? Here come three weeks of furniture ads following you everywhere.

This system is called segment targeting. Advertisers buy space to reach groups: 25- to 35-year-olds interested in sports, young parents, tech enthusiasts. It’s personalization, yes—but it remains general. You’re a profile among thousands sharing the same demographic or behavioral traits.

YouTube knows you’re 32 and watch fitness videos, so it puts you in the “active thirty-somethings” category and shows you protein powder ads. Facebook detects that you got engaged because you changed your relationship status, and suddenly you see hundreds of ads for wedding dresses and caterers.

It’s intrusive and sometimes annoying, but it’s still mass logic—you’re a data point in an ocean of similar data.

With ChatGPT, we enter another dimension

Now imagine you use ChatGPT every day. You talk to it. Really talk. You tell it things you wouldn’t even tell your loved ones. You ask it to help write a delicate email announcing your resignation it now knows you’re planning to leave your job. You ask for help writing a breakup message it knows your relationship is ending. You ask for advice on managing nighttime anxiety it knows you’re likely stressed and struggling with sleep. You ask about shared custody and financial implications of separation it understands you’re going through a divorce.

ChatGPT doesn’t put you in a category. It knows you.

Unlike Google, which guesses your intentions from searches, or Facebook, which analyzes your likes and friends, ChatGPT has direct access to your problems, doubts, and plans. You explain your symptoms, fears, and goals. You describe your financial situation, your career, your personal relationships. And this information is contextual, dense, and voluntary—you share it because you’re seeking help.

Hyper-personalization: from segment to individual

This is the real shift. We’re moving from category-based targeting to ultra-precise individual targeting. It’s no longer “you belong to the group of thirty-year-olds who like sports.” It’s “you are Sophie, 31, recently separated, suffering from chronic insomnia, earning $75K per year, with a history of impulsive purchases, currently seeking anxiety management solutions.”

Advertisers don’t access your raw conversations, of course. OpenAI has promised data isn’t sold. But advertising systems don’t need to read your messages to understand who you are. They only need to know what type of profile you match.

Take a concrete example. You talk to ChatGPT about your daily anxiety. You mention waking up in panic at night, difficulty concentrating, constant fatigue. ChatGPT gives meditation and breathing advice. Then, at the bottom of the response, an ad appears for Calm or Headspace.

On paper, it makes sense. You talk about stress, you’re shown a stress solution. But the algorithm knows much more:

You’re going through a difficult period (you mentioned your separation)
You have a good income (you asked investment advice)
You’re vulnerable (panic attacks, insomnia)
You tend to buy quick solutions (three wellness subscriptions recently)

The advertiser receives a signal: “premium profile, currently distressed, high conversion probability.” Targeting is no longer approximate. It’s surgical..

The problem: when help becomes sales

OpenAI says ads won’t influence ChatGPT’s answers, that everything will be transparent and clearly separated.

But is that really possible when economic incentives exist?

Reports in late 2025 revealed OpenAI employees discussed potentially giving “preferential treatment” to sponsored content. In other words, subtly steering responses toward paying advertisers. You ask “Which smartphone should I buy?” ChatGPT could answer objectively—or highlight partner brands. You ask how to sleep better it could suggest a sponsored paid app before mentioning free breathing techniques.

The risk isn’t aggressive sales. The risk is subtle drift, where your questions become commercial opportunities, and every conversation topic is analyzed for monetization potential. You’re no longer just asking for help you’re entering a sales funnel.

Why this is different from traditional targeting

On YouTube or Instagram, you know you’re in an advertising environment. It’s a public feed. Ads are part of the deal. You scroll, see an ad, and move on.

But ChatGPT is different. It’s not a feed it’s a conversation. A private exchange. You ask a question expecting useful, objective help. When an ad appears, it doesn’t just interrupt a feed. It enters a moment of trust. You shared something personal and something is sold back to you.

And the quality of information ChatGPT has about you goes far beyond what platforms collected before.

Facebook guesses your mood from interactions. ChatGPT you explain your mood directly, with context. That intimacy makes hyper-personalization both powerful and unsettling.

Claude bets on trust, ChatGPT on access

Anthropic turned this discomfort into a marketing argument. Their Super Bowl ads say: “No ads. Ever. You pay for the service, but you’re not the product.”

Sam Altman, on the other hand, argues advertising allows ChatGPT to be free for millions who can’t afford $20/month. The democratization argument: broader access is positive.

Both sides are right and wrong. Free access is good. But at what cost? If free AI becomes a sales machine, is it truly a win?

For Anthropic, “no ads” becomes a premium promise. You pay but you buy peace of mind.

What this means concretely for you

If you use ChatGPT for free, expect ads in coming weeks or months. They’ll appear below responses, clearly separated. OpenAI promises they’ll be discreet and relevant, and you’ll be able to hide them.

But beyond ads themselves, the principle changes. Each question triggers an evaluation of your potential value to advertisers. You become a commercial opportunity in real time.

Unlike cookies you can block, ChatGPT requires conversation. And conversation creates data.

OpenAI says you’ll be able to disable ad personalization but that only means random ads, not no ads. Tracking continues.

The end of AI neutrality

What’s happening isn’t just another advertising model. It’s the transformation of AI assistants into marketing tools.

Until now, chatbots presented themselves as neutral helpers. Advertising breaks that illusion. ChatGPT isn’t only there for you—it’s also there for advertisers.

Hyper-personalization isn’t just technical evolution—it’s a fundamental shift. You’re no longer one demographic profile among thousands. You’re an identifiable individual, with vulnerabilities, goals, and purchasing power.

The question is no longer:“Do you accept seeing ads?”

The question is:“Do you accept your intimacy becoming an advertising inventory?”

And that’s a question worth asking before clicking “I accept.”

Élisa Trousson