OpenAI Expands ChatGPT Ads to Health and Finance

A new analysis from SE Ranking reveals that ChatGPT displayed ads on 28.69% of commercial prompts within its healthcare category during a recent test. That figure is more than ten times higher than the 2.64% ad rate observed in Google’s AI Mode for the same vertical. The findings suggest that OpenAI’s current ad placement practices are far more permissive than its original launch materials indicate.

The shift traces back to April, when OpenAI quietly revised its ad placement guidelines. The company’s changelog frames the change as a “more precise approach,” and crucially, it removed the default block on ads appearing near medical, legal, and financial advice. Only conversations deemed sensitive remain off-limits.

OpenAI’s Current Ad Policies Explained

The official ad policies page, last refreshed on August 10, outlines two distinct layers of control: one governing where ads can appear, and another governing who is allowed to purchase them.

The placement layer prevents ads from showing in chats that OpenAI classifies as unsuitable, including political discussions. It also excludes what the company calls sensitive user contexts, a category that encompasses conversations about mental health and personal well-being.

The second layer concerns advertiser eligibility. Companies in finance, healthcare, and legal services must pass a manual, one-by-one review process before their ads are approved.

The healthcare list is broad. It includes health insurance, dental services, supplements, vision products, medical testing, and hospitals and urgent care facilities. The legal category, however, is treated differently. Ads for legal advice, court-related services, and similar offerings are prohibited, with the sole exception of general legal education. Political advertising remains entirely banned.

In July, OpenAI further clarified these distinctions by restructuring the health and finance sections, specifying exact categories and eligible markets.

Notably, the February launch announcement has not been updated to reflect these changes. While OpenAI has edited that original post three times (most recently on August 11 to add the UK, Mexico, Brazil, Japan, and South Korea to the pilot), the text still maintains that ads “are not eligible to appear near sensitive or regulated topics.”

What the SE Ranking Data Actually Shows

SE Ranking based its findings on 50,006 commercial prompts spanning 20 niches in the United States, collected on July 23. Of those, 12,974 prompts returned an ad, translating to a 25.94% overall ad rate.

The report groups results by niche but does not publish the exact prompts used. This means the healthcare figure reflects commercial queries categorized under that heading, not personal health conversations, which fall under OpenAI’s sensitive-context protections. The data also cannot confirm whether ads appeared in any conversations that those rules were designed to shield.

A similar caveat applies to the News & Politics category, which showed a 28.76% ad rate. SE Ranking notes that many prompts in that niche were subscription-related queries, such as newspaper or newsletter signups. The leading advertiser there was a direct mail company, indicating commercial media buys rather than political messaging.

The report also measured ad-to-prompt relevance. It found that 14.35% of ad impressions were no more related to their accompanying prompt than to a randomly selected one. That figure ranged from a low of 2.6% in the Pets category to a high of 54.2% in News & Politics.

SE Ranking offers ChatGPT and AI Mode ad tracking through its Competitive Research product, and the report concludes with a direct recommendation for that tool.

Why the AI Mode Comparison Has Limits

The 2.64% AI Mode figure originates from a separate SE Ranking report published on July 14, which was covered by SEJ last month. That study relied on 50,032 commercial keywords specifically engineered to trigger a text ad in Google, collected on June 30.

Because that keyword set was built for Google’s advertising system, comparing ChatGPT against it is not entirely apples-to-apples. The two datasets were also collected 23 days apart, a meaningful gap given how frequently both platforms update their ad policies.

The healthcare comparison, in particular, is constrained by sample size. AI Mode’s healthcare niche contained only 91 ads from 36 advertisers, making it the smallest category among the 20 niches in that report.

Why This Matters for Advertisers

The revised policy opens specific health and finance categories to case-by-case approval, and the enumerated list includes concrete product types. That level of detail makes the policy page a practical reference for evaluating whether a campaign might qualify.

Legal services, by contrast, are heading in the opposite direction. The summary language hints that approval could be possible, but the detailed rules explicitly forbid legal advice and court-related advertising. The launch announcement, which only gestures at a general restriction on ad placement near regulated topics, does not capture these nuances. Relying on that document for campaign planning would be a mistake.

The report’s headline numbers also should not be interpreted as evidence that OpenAI’s safeguards are failing. Ads appearing on healthcare-category commercial prompts does not necessarily mean the company’s placement rules were violated.

What to Watch Next

OpenAI has revised these ad rules four times in five months, with the most recent update on August 10 adding housing and job listings to the mix. Each iteration tends to shift which advertisers or categories qualify. A segment that is closed off when a media plan is drafted could easily be open by the time the campaign launches, so staying current with the policy page is essential.