How do you tell an ad from an answer in ChatGPT?

On January 16, 2026, OpenAI said it would soon start testing advertising in ChatGPT. The test is planned for logged-in adults in the United States on the Free and Go plans; paying Plus and Pro subscribers, and Business and Enterprise workspaces, will see none. As of this week, the ads are announced but not yet running, which makes this a good moment to learn what to look for before they appear.

The core idea is simple. ChatGPT's answer is the text the model writes in response to you. An ad is a paid placement from an advertiser. OpenAI says the two will be kept apart: ads at the bottom of an answer, clearly labeled, and separate from what it calls the organic answer, meaning the part nobody paid for.

Where ads will appear, and where they won't

According to OpenAI's plan, a test ad shows up below the answer when there is a sponsored product or service relevant to what you are discussing. You will be able to open an explanation of why you are seeing it, dismiss it, and tell OpenAI why.

Some places are off-limits during the test. No ads will be shown to accounts where the user says, or OpenAI predicts, they are under 18. Conversations about sensitive or regulated subjects, such as health, mental health or politics, will not carry ads. OpenAI also says advertisers will not see your conversations and that it will never sell your data to them.

A worked check

Imagine you are on the Free plan and ask:

I need a quiet dishwasher for a small apartment, under $700. What should I look for, and which models fit?

Here is how to read what comes back, step by step.

The answer itself also deserves the usual care. When ChatGPT's shopping research feature compiles a buyer's guide, OpenAI describes the results as organic, drawn from public retail sites, and it warns that prices and stock can still be wrong. Your own chats are not shared with retailers in that feature. That is a useful baseline: an unpaid answer can still be mistaken, so independence is not the same as accuracy.

One habit makes the line clearer. After any shopping answer, ask for the reasoning in a form you can check:

For each model you suggested, give the noise level, price range and one source I can open to confirm it. Put any model you are less sure about at the end.

Will ads change what it recommends?

OpenAI's stated answer is no. Its first advertising principle, which it calls answer independence, says ads do not influence what ChatGPT tells you and that answers are optimized for what is most helpful. The company also lists conversation privacy, the ability to turn off ad personalization and clear the data used for ads, and a promise that there will always be a way to use ChatGPT without seeing ads, including a paid plan with none.

Those are commitments, and the test is designed to show how they hold up in practice. CNBC notes that Sam Altman, OpenAI's chief executive, had expressed doubts about ads in earlier years because they could erode trust, before saying in November that the company would likely try them. Advertising also matters financially to a company with very large infrastructure commitments. Neither point means answers will be bent, but both are good reasons to keep watching rather than taking the promise as settled.

For a reader, the practical approach does not depend on trusting either side. Keep the visual line between the labeled ad and the answer in mind. Ask for sources and check the important ones. Compare a recommendation with at least one independent review before a significant purchase. And if you would rather not see ads at all, the paid tiers above Go are ad-free, while on the Free and Go plans you can switch off personalization so ads are not tailored to you.

If you notice an answer that seems to steer toward an advertiser, dismiss the ad, use the feedback option to say why, and keep a screenshot. During a test, that kind of report is exactly what shapes how the system develops.