Independent researcher Buchodi published an analysis on September 20, 2026 showing that OpenAI’s ads pixel uses a __obi cookie to link ChatGPT accounts to browsing and purchase activity on advertiser sites. On September 21, 2026, tech outlets in Europe and Asia amplified the findings and highlighted privacy and compliance concerns around the tracking design.
This article aggregates reporting from 5 news sources. The TL;DR is AI-generated from original reporting. Race to AGI's analysis provides editorial context on implications for AGI development.
This is the first detailed look at how a frontier AI lab is wiring classic adtech into a chat interface that people treat like a semi-private workspace. The technical findings matter because they show OpenAI using a long‑lived, cross‑site cookie to bridge ChatGPT accounts with browsing and purchase events on ordinary websites, even when users think they have opted out of marketing cookies. That moves ChatGPT from a neutral tool closer to a data‑hungry advertising platform, with all the regulatory and reputational baggage that implies.
For the race to AGI, the story is less about tracking mechanics and more about incentives. A lab that expects to burn hundreds of billions of dollars before turning cash‑flow positive will look for every monetization lever it can pull. If ad attribution becomes a core revenue stream, design decisions around models, logging and product UX will tilt toward engagement and profiling rather than minimal‑data architectures. That in turn raises the stakes for global privacy regulators, who now have concrete, packet‑level evidence to probe rather than abstract fears about AI. How regulators respond will shape whether future AGI‑class systems are financed by surveillance‑style ads, enterprise licensing, or something closer to public infrastructure.



