Health and finance advertisers get a route into ChatGPT ads, case by case
The rule most people still quote is the one OpenAI published in February: ads are "not eligible to appear near sensitive or regulated topics like health, mental health or politics". That stopped being the whole picture in April, and OpenAI never went back to update the February post.
The current position sits in OpenAI's ad policies, whose changelog dates the shift to version 1.1 in April 2026: "Medical, legal, and financial advice contexts are no longer categorically blocked from ads by default. Sensitive conversations and other prohibited contexts remain ineligible for ads."
Two separate layers decide whether an ad runs, and only one of them loosened. Placement governs which conversations can carry an ad at all, and that is where the regulated-advice block was lifted. Content governs who may advertise, and there the categories are still restricted.
In the US, OpenAI says it may approve health advertisers case by case for consumer medical devices and wearables, dental services, dietary supplements, disease-awareness campaigns, health insurance, hospitals and urgent care, medical testing, minimally invasive cosmetic procedures, and vision products and services.
Financial advertisers can be approved for auto loans, credit cards, credit monitoring, deposit accounts, financial planning, insurance, investment services, mortgages, personal loans and payment services. Proof of licensure may be required, and both categories are "generally prohibited" outside the US.
What did not change is worth stating plainly. Political content is still disallowed. Legal services still cannot advertise at all, even though legal-advice conversations became eligible to carry ads, which is the one place the two layers point in opposite directions. Sensitive user contexts, including suicide and self-harm, remain off limits.
Search Engine Journal, which set the change out on 12th August, makes the practical point: OpenAI has revised these policies four times in five months, so any campaign assumption more than a few weeks old is worth re-reading.
Correction, 15th Aug 2026. Our ads tracker below said the pilot's rules were "unchanged since launch" and that no ads run on health, mental-health or politics topics. That stopped being accurate in April 2026, so it was wrong when this page was built on 13th August. The tracker and the ads FAQ now describe the current two-layer policy, and the sources are unchanged.