OpenAI's current advertising policy no longer treats every health and financial advertiser under one blanket exclusion. In the United States, selected advertisers in those categories may be accepted through manual review when licensing and content conditions are met. A separate protection governs placement: even an approved advertiser cannot appear beside personal health, mental-health, political, or other sensitive conversations.
The practical consequence is that marketing teams must separate two questions. “Can this advertiser enter the programme?” is not the same as “Beside which conversation can this ad appear?” Advertiser eligibility, creative review, and landing-page review form one layer. User context and brand safety form another. Approval in the first layer does not override the second.
What changed, and what does the policy say now?
OpenAI's advertising policy says supported consumer categories may expand during the test. The current text allows approved advertisers in finance, healthcare, and medicine to be considered through manual, case-by-case review. This is not a general permission. Product type, market, licence, claim, and destination are evaluated together.
In US financial services, eligible areas may include auto lending, cards, deposit accounts, planning, insurance, investment services, mortgages, personal loans, and payments. Credit repair, debt settlement, and alternative investments remain prohibited. Financial-services advertising outside the US is generally not allowed.
For approved US health advertisers, reviewable areas include selected consumer medical devices, wearables, dental services, insurance, hospitals, urgent care, testing, and diagnostics. Unsupported treatment claims, experimental therapies, high-risk procedures, and products that exploit vulnerability are prohibited. A general wellness product is not treated the same as a disease-treatment claim.
This text-free Fark Studio illustration presents advertiser eligibility and the sensitive-conversation placement barrier as two separate gates. It is not a ChatGPT interface.
Important boundaries for legal, political, and health contexts
Legal-service advertising is generally prohibited, while general legal education or test preparation may be considered when no service is offered. Political advertising is prohibited. Gambling, tobacco, adult content, misleading claims, and fraud are also unsupported. Approval for a health or finance advertiser does not create unlimited freedom for creative claims.
Placement policy is a separate control. OpenAI says ads do not appear in personal health and mental-health conversations, emotionally reliant interactions, sensitive user journeys, political content, fraud, harm, and other brand-unsafe contexts. An approved health-insurance advertiser therefore cannot appear next to a user's personal diagnostic discussion.
That distinction can also affect reporting. Lower-than-expected reach for an approved campaign may not be only a bid or creative problem. The inventory of suitable conversations may be intentionally narrow. OpenAI has not yet explained how much detail about contextual filtering will appear in campaign reporting.
Who is affected, and what is the position for Türkiye?
Banks, insurers, payment companies, health-technology brands, hospitals, clinics, wearable-device makers, and their agencies should monitor the development. But today's policy sets a clear market boundary for an advertiser based in Türkiye: case-by-case health and financial eligibility is focused on the US, while those services are generally prohibited outside the US.
The broader ChatGPT ads pilot is expanding beyond the US into markets including Canada, Australia, New Zealand, the United Kingdom, Mexico, Brazil, Japan, and South Korea. That does not mean regulated-category advertising has opened in Türkiye. No Türkiye launch date appears in the official material.
Teams can still prepare. A global company needs to map the US licensed entity, advertiser, destination, and target market. Local teams should avoid placing a firm budget in the plan based on an assumption of imminent Türkiye availability.
What is confirmed and what remains uncertain?
Confirmed points include the separation of eligibility and placement, possible manual approval for selected US health and financial advertisers, the exclusion of ads from personal sensitive conversations, the prohibition on political ads, and the requirement for end-to-end consistency between creative and destination.
Unknowns include manual-review timing, minimum spend, a Türkiye rollout, the level of contextual-filter reporting, and the detailed appeal path after a rejection. OpenAI also says its category list may evolve as the programme develops.
What should brands in Türkiye do now?
1. Verify market eligibility first. Possible US approval does not open a Türkiye-targeted campaign. 2. Match the advertising entity with the relevant licence. The landing-page company and applicant need a clear responsibility chain. 3. Connect every health and finance claim to an evidence file. Avoid guarantees, certain outcomes, and unsupported comparisons. 4. Audit category consistency from creative through destination. An eligible-looking ad must not lead to a prohibited product. 5. Minimise sensitive-data use and do not build targeting plans that assume advertiser access to conversation content. 6. Add unsuitable conversation contexts to the brand-safety brief and explain that advertiser approval does not produce unrestricted placement. 7. Bring legal, compliance, media, and web teams into one checklist before applying. 8. Model a constrained-inventory scenario in performance forecasts instead of comparing reach directly with a large social platform. 9. Do not trust brokers promising guaranteed access before official Türkiye eligibility is announced.
This Fark Studio workflow combines licensing, claims, destination, market, and human review. It is not an official OpenAI application screen.
Where should teams wait?
Wait before committing Türkiye budget, forecasting reach, or assuming approval for a regulated category. Do not sign media-buying commitments until the official application and market eligibility are visible. In regulated sectors, do not rely on an industry headline alone for policy interpretation.
Evidence and destination quality should not wait. Disorganised licensing, company information, claim support, privacy notices, and product scope will make review harder when a market opens.
Fark Studio perspective
Safety in ChatGPT advertising is not merely a prohibited-word list. Advertiser, content, destination, market, and conversation context pass through separate gates. A regulated brand should first be able to prove which promise is allowed in which market, then pursue speed.
Performance marketing, digital marketing, and web design teams should build the approval file together. If you need a market, claims, and landing-page audit for regulated advertising, plan an advertising-eligibility review with Fark Studio.
Sources
OpenAI, Ad policies, updated July 15, 2026; eligibility, placement, and category boundaries.
OpenAI Help Center, Ads in ChatGPT, accessed August 12, 2026; sensitive-conversation placement and regulated-advertiser guidance.
OpenAI, Testing ads in ChatGPT, updated May 7, 2026; official pilot-market coverage.
Search Engine Journal, OpenAI allows some health and finance ads in ChatGPT, August 12, 2026; current industry summary of the policy.
