The Pulse
OpenAI Tightens ChatGPT Ad Rules, Reserves Broad Veto Power
OpenAI updated its ChatGPT advertising policies on September 10, adding language that lets the company reject ads that conflict with its business interests or competitive position. The revised rules also detail restrictions on political, he
AI.info Team ·
OpenAI says ChatGPT ads will not influence the answers users receive. Its revised advertising policy gives the company wide discretion to reject an ad anyway, including when it conflicts with OpenAI’s advertising principles, business interests or competitive position.
The policy page, updated September 10, marks version 1.6 of OpenAI’s advertising rules. The change adds the explicit business-interest and competitive-position language to the company’s existing right to decline, restrict or remove advertisers and ad content.
That creates a clear tension in OpenAI’s approach: the company presents advertising as a carefully separated layer around ChatGPT’s answers, while retaining broad authority over which commercial messages can appear and which advertisers can participate.
OpenAI’s new sentence changes the policy’s reach
The revised policy says OpenAI may refuse to display advertising or website links “for any reason,” including conflicts with its advertising principles, business interests or competitive position. The page does not define how OpenAI will apply the competitive-position test or whether it could cover ads from companies that compete directly with OpenAI’s products.
OpenAI’s updated policy also describes a three-stage review process covering advertisers, ad creative and landing pages, and the conversations in which ads appear. The company says automated systems, including large language models and classifiers, conduct most reviews, with human oversight and additional review for higher-risk cases.
Ads that cannot be evaluated by OpenAI’s systems are not eligible to run. The company says it may reject or remove an ad, limit its delivery, require changes, restrict an advertiser’s access or suspend an account after severe or repeated violations.
Political and sensitive conversations remain off limits
OpenAI prohibits ads near conversations involving politics, child safety, cyber abuse, weapons, terrorism, suicide or self-harm, misinformation, fraud, graphic violence and regulated goods such as gambling or tobacco. It also bars ads near emotionally reliant interactions, mental-health discussions, personal-health conversations and other sensitive user journeys.
The policy separately prohibits political advertising, including messages supporting or opposing candidates, parties, elections, voting, public officials, government action or contested social issues. Advertising that meaningfully references abortion, immigration, race, religion or gender identity is also barred when it treats those subjects as political or socially contested matters.
OpenAI says the early advertising program focuses mainly on household and consumer products, local services, travel and entertainment, and digital products or education. Other categories may be approved gradually, but the company applies additional review to financial services, healthcare and legal services.
Financial, health and legal ads get narrow U.S. exceptions
The updated rules allow certain U.S. financial advertisers to apply on a case-by-case basis. Eligible categories include credit cards, deposit accounts, insurance, mortgages, personal loans, payment services, investment services, financial planning, credit monitoring, and auto loans or leasing.
OpenAI continues to prohibit credit-repair, debt-settlement and debt-assistance advertising, along with alternative investments such as bullion. Financial-service advertising outside the United States is generally prohibited under the policy.
Health advertising receives a similar treatment. OpenAI may approve U.S. ads for medical devices, wearables, dental services, dietary supplements, hospitals, urgent care, diagnostic services, health insurance and vision products, provided the advertiser meets the company’s review and compliance requirements.
Unsupported treatment claims, experimental therapies, unsafe medical products, invasive high-risk procedures and products that exploit body-image concerns remain prohibited. Legal advertising is allowed only in the United States and only when the advertiser is licensed in the jurisdiction where the ad appears.
The policy follows a wider ChatGPT ads rollout
OpenAI began testing ads in the United States in February for users on its Free and Go plans. Plus, Pro, Business, Enterprise and Edu accounts do not receive ads, according to the company’s Help Center guidance.
OpenAI expanded ChatGPT ads to 31 European markets in August, including Germany, France, Spain, Italy, Sweden, Norway, Denmark, the Netherlands and Austria. The company said advertisers would initially work through OpenAI’s sales team and partners, with self-service access through Ads Manager added across those markets on August 31.
The European rollout also broadened the commercial system beyond basic campaign buying. OpenAI says it now supports conversion optimization, geographic targeting, custom audiences and measurement through the OpenAI Pixel, Conversions API and third-party integrations.
OpenAI said tens of thousands of marketers had advertised on ChatGPT by August. The September policy revision gives that growing advertiser base a more detailed set of restrictions, while leaving the company significant room to decide which businesses and campaigns it will accept.
“You need to trust that ChatGPT’s responses are driven by what’s objectively useful, never by advertising.”
Fidji Simo, then CEO of Applications, OpenAI
OpenAI promises separation, but controls the whole pipeline
OpenAI’s advertising principles state that ads are clearly labeled, displayed separately from ChatGPT’s answers and unable to shape or rank those answers. The company also says advertisers do not receive users’ chats, memories, names, email addresses, precise locations, IP addresses or sensitive information.
Users can turn off ad personalization, clear advertising data and dismiss or report individual ads. OpenAI says a free no-ads option is also available in some regions, though it carries lower message limits and reduced access to features such as image generation or deep research.
Those controls address how ads reach users. The new policy addresses who gets to reach them in the first place. OpenAI’s September 10 update makes clear that the final decision belongs to OpenAI, even when an ad meets the published category and safety requirements.