How ChatGPT serves ads
buchodi.com
April 28, 2026
3 min read
🔥🔥🔥🔥🔥
61/100
Summary
OpenAI's ad platform consists of a backend that injects structured single_advertiser_ad_unit objects into the ChatGPT conversation stream. A tracking SDK called OAIQ operates in the visitor's browser to report product views, linking the ad serving and tracking through Fernet-encrypted click tokens.
Key Takeaways
- OpenAI's ad platform integrates a backend that injects structured ad units into ChatGPT's conversation stream while the model responds.
- A tracking SDK called OAIQ runs in the user's browser to report product views back to OpenAI, linking ads and user interactions through Fernet-encrypted click tokens.
- Ads delivered in ChatGPT are contextually targeted based on the conversation topic, with different advertisers appearing across various discussions.
- Each ad includes four distinct Fernet-encrypted tokens for tracking and integrity checks, ensuring secure attribution and monitoring of user interactions.
Community Sentiment
Positives
- The introduction of ads in the free tier and new ad-supported plan could help sustain OpenAI's services while maintaining access for users who cannot afford subscriptions.
- The revenue from ads on the free tier may justify its existence, potentially allowing OpenAI to continue offering valuable AI tools without significant cost to users.
Concerns
- The shift towards ad-supported models raises concerns about the potential for adversarial content and the integrity of AI-generated responses.
- Injecting ads into model outputs could breach user trust, leading to long-term detrimental effects on the perceived reliability of AI systems.
- There is a fear that ads could become indistinguishable from actual content, undermining the quality of information provided by AI.