Key Points
- Ads shown only in Free and Go tiers
- ChatGPT answers remain independent
- Advertisers receive aggregate metrics only
Summary
OpenAI is piloting ads in ChatGPT to help fund Free and Go tiers while preserving answer independence, user privacy, and meaningful user controls. The initial U.S. test (begun Feb 9, 2026) targets logged-in adult users on Free and Go only; Pro, Business, Enterprise, and Education tiers are ad-free. Ads are clearly labeled, do not change model outputs, and advertisers receive only aggregate performance metrics.
Key Points
- Scope & timeline
- Started in the U.S.; planned expansions: Canada, Australia, New Zealand (announced Mar 26, 2026), and UK, Mexico, Brazil, Japan, South Korea (announced May 7, 2026).
- Businesses can sign up at openai.com/advertisers for participation updates.
- Targeting & ad selection
- Ads are matched to conversation topic, past chats, and past ad interactions to improve relevance.
- Multiple advertisers are ranked and the most relevant ad is shown first.
- Privacy & safety
- Advertisers do not get access to chats, chat history, memories, or personal details—only aggregate metrics (views, clicks).
- Ads are excluded for under-18 accounts and near sensitive or regulated topics (health, mental health, politics) during the test.
- User control
- Users can dismiss ads, provide feedback, see why an ad was shown, delete ad data, and manage personalization settings.
- Option to upgrade to paid tiers or opt out in Free tier (with fewer daily messages) to avoid ads.
- Engineering implications
- Implement and maintain: ad labeling and visual separation, relevance ranking, privacy-preserving telemetry, age detection and filtering, sensitive-topic classifiers, opt-out and data-deletion APIs, and advertiser guardrails to reduce fraud and harmful content.
- Goals
- Learn from real-world usage to iterate on relevance, safety, and controls while keeping ChatGPT answers independent and user conversations private.