ANKARA 

The integration of personalized ads into artificial intelligence (AI) chatbots and search systems is reshaping how users access information and how digital advertising functions.

OpenAI said on Feb. 9 that it has begun testing ads on ChatGPT in the US, with placements determined by factors such as chat topics, past conversations, and previous ad interactions.

A user discussing or searching for recipes could see ads related to meal kits or grocery delivery, while users can manage their ad preferences.

OpenAI said advertisers cannot access chat content or personal data, except for reportedly anonymous and aggregated performance data such as total impressions and click counts. However, legal and ethical concerns around such practices remain unresolved.

In traditional search engines, queries generate sponsored links placed at the top of results, allowing users to recognize advertisements and decide whether to engage. In contrast, next-generation AI systems use large language models (LLMs) to deliver a single, comprehensive response, where advertising may be embedded within the answer itself.

This shift is altering both the structure of digital advertising and how users interact with it.

Koen Pauwels, a marketing professor at Boston-based Northeastern University and editor-in-chief of the International Journal of Research in Marketing (IJRM), told Anadolu that integrating informative content with paid placements makes it harder for users to distinguish between neutral responses and commercial influence.

Pauwels said AI chatbots generate richer datasets than traditional search engines by building behavioral profiles that capture not only user intent but also emotional context.

For example, when users seek advice on chronic pain, compare mortgage options, or plan travel, they may reveal personal vulnerabilities that AI systems can use to refine ad targeting with greater precision.

Pauwels urged users to be skeptic of platforms claiming their AI systems are not under commercial influences, such as Anthropic, which ran an ad during the annual Super Bowl game mocking OpenAI’s ChatGPT planning to embed ads in AI responses, while Anthropic’s Claude claims not to have integrated ads.

He said that given how LLMs operate, there is often no clear boundary between information and paid content, with distinctions depending on system design, search architecture, and disclosure policies.

Meanwhile, the long-term trajectory of the Al industry points to a future where autonomous Al agents will take over tasks like booking flights, comparing insurance, and ordering groceries on behalf of the user, which could mean that the fundamental nature of digital marketing will shift since brands will have to convince algorithms for their products to be chosen, rather than the human customer.

Pauwels added that this could lead to efficiency but also the loss of conscious choice for the customer, stripping away the human choice.

Pauwels said the digital advertising market, currently dominated by Google, could consolidate around a small number of AI platforms if such agents replace traditional browsing, increasing their influence over advertisers, publishers, and user behavior.

He noted that while the internet has traditionally offered free access to information in exchange for user data, AI systems are now operating within a more intimate cognitive space between humans and machines.

Pauwels said AI platforms collect extensive data, treating each prompt as a data point and each session as an opportunity to refine behavioral profiles.

He said these systems can model user psychology and deliver highly tailored responses aligned with individual decision-making patterns, which may lead to outcomes that border on exploitation rather than persuasion, particularly for vulnerable users.

He urged for the creation of a clearer social contract among users, regulators, and platforms over what is acceptable in exchange for personal data

*Writing by Emir Yildirim in Istanbul