Chatbots and ads: what will "free" intelligence cost us?

Remember the blessed days when Google was just a search bar, without five ad links on top? With large language models, this phase of "pure love" and selfless help looks set to end much sooner than we expected. While venture capitalists showered startups with billions, we got used to the planet's most powerful intelligence being a click away and not trying to push a new blend of coffee on us. But the party on someone else's dime is winding down, and people in sharp suits with slide decks about traffic monetization are taking the stage. The gap between the cost of generating a single answer and the revenue from free users has become too large to ignore.

Perplexity was the first to jump into the commercial pool. The AI search service no longer just hints at it; it states outright that it is introducing sponsored questions. It works elegantly and insidiously: after the main answer, the system suggests "follow-up questions," some of which are paid for by brands. If you ask about the best running shoes, don't be surprised if one of the suggested conversation paths leads you straight into the arms of a specific manufacturer. This is no longer just search; it is guided navigation through the consumer world, where the line between recommendation and advertising blurs until it is invisible.

Google is in the most difficult position. For the Mountain View company, putting AI into search is both a lifeline and an existential threat to its core business. If Gemini gives the perfect answer right away, why would users click on ad links? The solution turned out to be predictable: ads seep into the so-called AI Overviews. Blocks labeled "Sponsored" now appear directly above or inside the generated text. Google spent too long building its advertising empire to let some neural-network progress break the supply chain of clicks to advertisers.

Microsoft is keeping pace with Copilot. Because the service is tightly integrated with the Bing search engine, ad links and banners inside the chat have been the norm almost since day one. Users have already started noticing that the bot sometimes pushes specific services or products too insistently, disguising them as useful links in its sources. This sets a dangerous precedent: we have grown used to trusting a "smart" assistant more than search results, and that trust is very easy to abuse for the sake of quarterly reports.

OpenAI's behavior remains the most interesting. Sam Altman long said publicly that he dislikes the advertising model and prefers subscriptions. Yet the reality is that ChatGPT demands colossal computing power, and hundreds of millions of free users generate only losses. Recent news about hiring specialists from the ad departments of Google and Meta* speaks for itself. OpenAI will most likely choose the path of "premium purity": those who pay $20 a month will talk to pure intelligence, while everyone else becomes a target audience for contextual offers.

The problem with ads in chatbots runs much deeper than annoying banners. In ordinary search, we see clearly: here is the ad, there is the result. In a conversation with AI, the ad message becomes part of the narrative. A bot can weave a brand mention into its advice so organically that you won't even realize where objectivity ended and marketing began. That turns a personal helper into the perfect salesperson, one who knows everything about you and knows how to persuade. We are entering an era in which freedom from manipulation in answers will carry its own separate tax.

Bottom line: Ads in AI are inevitable, because the bills for electricity and H100 chips won't pay themselves. Are you ready to trust a bot's advice, knowing its "opinion" may be bought?

*Meta is designated an extremist organization and banned in Russia.