Ad Break in ChatGPT: US Senate Questions OpenAI's Integrity

In the coming weeks, OpenAI will start testing ads in the free version of ChatGPT: "sponsored" products and services will appear at the bottom of the chat window. In response, US Senator Ed Markey (D-Mass.) sent letters to the heads of OpenAI, Anthropic, Google, Meta, Microsoft, Snap, and Elon Musk's xAI, demanding they explain how advertising in AI chatbots will affect consumer protection, privacy, and the safety of underage users.

What's changing in ChatGPT

According to The Verge, OpenAI will begin testing ads for free ChatGPT users in the coming weeks. The format: "sponsored" products and services placed at the bottom of a conversation with the chatbot. The company says the ads will be relevant to the context of a specific conversation, but that it will not use users' personal data to train models or target ads.

For OpenAI, this is effectively a first step toward the internet's classic advertising model. Its hundreds of millions of free users are expensive to serve, and running models like GPT-4o requires enormous computing power. The line between a "contextual suggestion" and "paid promotion" is blurry here: if the bot sees that a user is asking about sneakers and immediately shows a link to a specific brand, it is already using the conversation's context for advertising purposes.

Why Senator Markey is alarmed

In letters addressed to the companies' CEOs, Markey writes that introducing ads into AI chatbots "raises serious concerns about consumer protection, privacy, and the safety of young users" (quoted from the document published on markey.senate.gov). The senator's reasoning: teenagers tend to trust AI models more than ordinary search engines, so hidden commercial influence could distort how they perceive information. The worry is that models will start optimizing not for answer accuracy but for the likelihood of a click on a sponsored link. In other words, the algorithm would quietly nudge users toward a purchase disguised as "smart advice," without the explicit signals that classic search results provide with an "ad" label.

Who else received letters from the senator

Markey's inquiry did not stop at OpenAI. Anthropic, Google, Meta*, Microsoft, Snap, and xAI received similar letters demanding explanations. Formally, the senator wants to know whether these companies plan to introduce ads in their own AI chatbots and what safeguards they have for consumers and minors. The list includes seven of the market's biggest players at once, from assistant developers (Anthropic, xAI) to owners of messaging apps and social networks (Meta, Snap). That shows Washington sees advertising in AI not as an OpenAI-specific issue but as a systemic risk for the entire industry.

Why the timing is risky for OpenAI

For OpenAI, advertising is a forced move: running the free version of ChatGPT for hundreds of millions of users has become a serious financial burden, and the Plus subscription alone does not cover it. But the timing looks poor. AI safety and ethics are already on regulators' agenda, and trying to turn a chatbot into an ad platform gives them more reasons for oversight. In effect, the industry now faces a fork in the road: either move to a transparent monetization model with clear rules for displaying ads, or risk digital assistants working not only for users but also for advertisers.

When will ads appear in ChatGPT?

OpenAI will begin testing ads for free ChatGPT users in the coming weeks. They will appear as "sponsored" product and service offers at the bottom of the chat window and, according to the company, will be relevant to the context of a specific conversation.

Which companies received letters from Senator Markey?

Senator Ed Markey sent letters demanding explanations to seven companies: OpenAI, Anthropic, Google, Meta*, Microsoft, Snap, and xAI. He wants to know whether they plan to introduce ads in their AI chatbots and how that would affect consumer protection and the safety of minors.

Will OpenAI use users' personal data to show ads?

According to OpenAI, no. The company says it will not use users' personal data either to train models or to target ads, although the ads themselves will be selected based on relevance to the context of the current conversation.

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