OpenAI tests ChatGPT ads with guardrails; brands get sponsored slots below answers

New Delhi: OpenAI has begun testing ads in ChatGPT in the US, and is introducing a clear trade-off for Free users: stay on Free with ads, or opt out of ads on Free “in exchange for fewer daily free messages”.

The test applies to logged-in adult users on the Free and Go subscription tiers. OpenAI said Plus, Pro, Business, Enterprise and Education tiers will not carry ads.

OpenAI has positioned “answer independence” as the core guardrail. It said ads “do not influence” ChatGPT’s answers and will be “clearly labeled as sponsored” and “visually separated” from the organic response.

The company also disclosed how ads will be picked during the test. It will match advertiser-submitted ads with the topic of the conversation, a user’s past chats and past ad interactions. If more than one advertiser is eligible, OpenAI said it will show the most relevant ad first.

OpenAI said advertisers will not have access to chats, chat history, memories or personal details. Advertisers will only receive aggregate performance information such as views or clicks.

On brand safety, OpenAI said it will not show ads in accounts where the user says they are under 18, or where OpenAI predicts they are under 18. It also said ads are not eligible to appear near sensitive or regulated topics such as health, mental health or politics during the test.

OpenAI added that as it scales the programme, it plans guardrails to prevent narrow ad targeting, and said it will be “deliberate” about which advertisers are allowed in, while building protections against scams and misleading ads.

OpenAI said users will get controls to dismiss ads, share feedback, understand why a particular ad is being shown, delete ad data “with one tap”, and manage ad personalisation at any time.

OpenAI said ChatGPT is used by “hundreds of millions” of people for learning, work and everyday decisions. It said keeping Free and Go “fast and reliable” requires significant infrastructure and ongoing investment, and ads help fund broader access through “higher quality free and low cost options”.

As per the news reports, participation in the early advertising programme will require a minimum spend of $200,000, and the company plans to track basic metrics such as clicks and impressions while evaluating additional measurement tools as the testing progresses.

OpenAI said the focus of the test is learning and feedback before it expands. It said it will evolve its advertising programme over time to support additional formats, objectives and buying models, and has opened an advertiser interest sign-up page.

OpenAI has set up an “ads integrity” team with a focus on safeguards around advertiser verification, brand safety and user trust.

Separately, OpenAI’s move towards ads has already triggered pushback from rivals. Recently, Anthropic released a series of commercials mocking the idea of sponsored content appearing inside chatbot responses, including a film in which a ChatGPT-like assistant shifts from fitness advice to promoting insoles.

OpenAI CEO Sam Altman hit back on X, calling the ads “dishonest” and saying OpenAI’s ad guidelines would not allow such in-response placements. “We would obviously never run ads in the way Anthropic depicts them. We are not stupid and we know our users would reject that,” Altman wrote, while also conceding the ads were “funny” and that he “laughed.”