ChatGPT Ads Test Live — Free/Go Monetization Without Touching Paid Trust Tiers

January set principles; February turns them on. OpenAI’s proprietary needle-thread: monetize Free/Go with labeled ads while keeping high-ARPU tiers pristine—so subscription trust and ad ARPU can coexist in one product.

OpenAI started testing advertisements in ChatGPT on February 9. Scope: U.S. logged-in adult users on Free and Go only. Pro, Business, Enterprise, Education remain ad-free. Placement: clearly labeled, visually separated below organic answers. Targeting uses conversation topics/history but must not alter responses (company principles). Advertisers receive aggregate performance data only—no individual chats. Controls: dismiss, feedback, delete ad data, manage personalization, opt out of personalized ads (with trade-offs such as fewer free messages). Test bounds: no ads for users under 18; ads not eligible near sensitive topics (health, mental health, politics) during the test. Goal stated: support broader access on lower-cost tiers while funding infra/model work; start small before wider rollout. Later updates on the same post note phased expansions—the initial test launch is dated February 9.

Claims vs checks

Tier scope, placement, and privacy rules are OpenAI primary (“Testing ads in ChatGPT”). Whether ads stay non-influential in practice is a behavioral claim—not independently audited here. Help Center / TechCrunch confirm U.S. start and tier restrictions.

Limits

• Test ≠ global GA; inventory and topic blocks can change.
• Opt-out personalized ads may reduce free-tier quotas—read the control copy.
• Aggregate advertiser metrics still create optimization pressure over time.

Sources

• OpenAI: “Testing ads in ChatGPT” (February 9, 2026). Primary.
• Help Center and contemporaneous reporting (TechCrunch, etc.) on U.S. start, tiers, placement/privacy.
• Related: January 16 advertising-principles post for timeline.