--- title: ChatGPT now knows what you do on other websites via ad... description: OpenAI's ad measurement system sets a cookie called __obi, scoped to .openai.com with SameSite=None, that gets sent back to OpenAI whenever a user visits any... canonical: https://daily.dev/posts/chatgpt-now-knows-what-you-do-on-other-websites-via-ad-collector-6ydoidhyl twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: ChatGPT now knows what you do on other websites via ad collector | daily.dev og:description: OpenAI's ad measurement system sets a cookie called __obi, scoped to .openai.com with SameSite=None, that gets sent back to OpenAI whenever a user visits any... og:url: https://daily.dev/posts/chatgpt-now-knows-what-you-do-on-other-websites-via-ad-collector-6ydoidhyl og:image: https://api.daily.dev/og/posts/6YDOidHyL.png og:image:alt: ChatGPT now knows what you do on other websites via ad collector og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # ChatGPT now knows what you do on other websites via ad collector **[Lobsters](https://daily.dev/sources/lobsters)** · 7 min read · 0 upvotes · 0 comments ## Summary OpenAI's ad measurement system sets a cookie called __obi, scoped to .openai.com with SameSite=None, that gets sent back to OpenAI whenever a user visits any site running OpenAI's advertiser pixel. The investigation reproduced the mechanism on a phone, traced the identifier from a signed JWT issued by ChatGPT's backend through a cross-site sync request, and observed it firing on sites like Chewy, Wayfair, ThriftBooks, Eventbrite, HelloFresh, Coursera and SeatGeek. The pixel also scrapes emails, phone numbers, and other identity fields from advertiser pages via hijacked tag-manager data layers, with scraped identity outnumbering advertiser-supplied identity in observed traffic. The mechanism works even for logged-out users via a persistent anonymous identifier, and only affects Chrome (not Safari or iOS browsers due to third-party cookie blocking). OpenAI's cookie policy classifies __obi as an analytics cookie despite it functioning as a cross-site marketing/ad-tracking identifier tied to ChatGPT accounts; OpenAI did not answer direct questions about this classification when contacted. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### What is the __obi cookie that OpenAI sets and why does it get sent to other websites? __obi is a cookie set by OpenAI's ad collector at bzr.openai.com, scoped to .openai.com with SameSite=None and Secure, which allows it to be attached to cross-site requests. It is created from a signed JWT issued by ChatGPT's backend that binds the identifier to a ChatGPT account or anonymous session, then gets sent to OpenAI whenever a page running OpenAI's advertiser conversion pixel loads, letting OpenAI link browsing on that site to a ChatGPT identity. _Anyone auditing third-party ad pixels for cross-site identity leaks can track findings like this on daily.dev._ ### Does OpenAI's ad tracking cookie __obi still work if a ChatGPT user is not logged in? Yes, the mechanism works for logged-out users. Across 932 decoded sync tokens, 736 carried subject_type account_user and 196 carried anonymous, and the anonymous subject proved just as stable as the account-linked one, persisting on a device for at least 27 days, meaning cross-site tracking continues even without an active login. _Developers weighing consent and identifier persistence in their own products can follow analysis like this on daily.dev._ ### Does OpenAI's advertiser pixel collect personal data like email and phone number from advertiser websites? Yes, the pixel SDK scrapes identity data directly from advertiser pages, not just from data advertisers deliberately supply. It hijacks the tag-manager data layer (window.dataLayer, adobeDataLayer, renamed GTM layers) to pull emails and phone numbers, and in observed traffic scraped identity events outnumbered advertiser-supplied ones 685 to 255; email, phone, and names are SHA-256 hashed, while country, region, city and postal code are sent unhashed. _Teams reviewing what their ad-pixel integrations actually collect can keep up with findings like this on daily.dev._ ## Community take How the wider developer community reacted, aggregated from 2 discussions and 708 comments across lobsters, hackernews (as of 2026-09-21). **TL;DR:** The discussion is dominated by disgust at OpenAI running standard adtech tracking inside a chat product, with debate spiraling into broader arguments about surveillance capitalism, corporate accountability, and (in a long tangent) whether LLMs are actually 'intelligent'. **Sentiment:** 5% positive · 15% mixed · 80% skeptical **The case for** - A few argue this is just normal ad-industry data sharing that improves ad relevance and isn't 'spying' in a meaningful sense. - Some note third-party cookie tracking is a known, decades-old mechanism rather than something uniquely sinister to this product. **The pushback** - Many feel visceral discomfort that an AI chat product, which people trust with intimate queries, is now also a cross-site ad-tracking vehicle. - Commenters argue this reflects a broader failure of private enterprise/capitalism to self-regulate around surveillance. - Several blame individual engineers and companies for building these systems and normalizing 'spyware' language like 'standard adtech'. - Some see this as part of a pattern where regulation (e.g. EU rules) is the only real check on such tracking, though its effectiveness is contested. **By community** - hackernews (heated): Strong initial disgust at the tracking mechanism quickly fragments into heated, tangential debates about capitalism, propaganda, AI intelligence, and regulation, with only a minority defending the practice as ordinary adtech. **Hottest debate:** Whether this is a uniquely alarming abuse of trust by an AI product or just normal, longstanding adtech behavior that shouldn't surprise anyone. **Open questions** - Why did OpenAI classify the cookie as 'analytics' rather than disclosing its cross-site marketing function, and will regulators or OpenAI address that discrepancy? - Will browser/platform changes (e.g. further third-party cookie restrictions) meaningfully curb this kind of tracking going forward? **Highlights** > To me, this quote just about sums it up: > The mechanism is standard adtech. What has no precedent is running it on an AI chat product. As someone who has been well aware of this mechanism for quite some time, I still feel icky anytime I re-read the details of it. What a time to be alive. > — [mavsman on hackernews · 6 comments](https://news.ycombinator.com/item?id=49777671) > Whenever I say something like “that’s a cool feature, but to do it you would have to build spyware”, everyone else is just like “the cat is out of the bag ¯\_(ツ)_/¯”. (I don’t build spyware, or work on projects that do). It blows my mind that people don’t care about the world they are building with this stuff. It’s a real tragedy of the commons. People see these collaborators from different wars and regimes and think “I’d stand up against the bad guy”… well I’ve got news for you if you build spyware, you are not the person you think you are. > — [clickety\_clack on hackernews · 6 comments](https://news.ycombinator.com/item?id=49777865) > Your issue is you are defining spyware too broadly. This is not spying on users, but rather 2 companies partnering and sharing data to result in either a better ads system or better understanding on how ads are performing. This is a positive value to society and the commons. Wasting space in a site or app with an ad that won't convert is the true tragedy of the commons. It's a waste of time and money for all parties involved. > — [charcircuit on hackernews · 2 comments](https://news.ycombinator.com/item?id=49778547) > "standard adtech" I hate stuff like this. Sometimes euphemisms are kind, like "senior citizen" instead of "old person". But this is an attempt to normalize bad behavior that is really quite terrible for society. > — [m463 on hackernews](https://news.ycombinator.com/item?id=49782981) > Yeah, this is just standard third-party cookie functionality, which has always been sketchy. It honestly seems like it was only possible by accident; browsers have long prevented sites from reading cookies from other domains, but it seems like the people working on early specs might not have considered the ramifications of being able to set cookies for domains other than your own. A couple decades ago it might have seemed like no one would have any reason to set a cookie they couldn't read. > — [saghm on hackernews · 1 comments](https://news.ycombinator.com/item?id=49782279) **Source threads** - [lobsters](https://lobste.rs/s/jbnmj9/chatgpt_now_knows_what_you_do_on_other) · 25 points · 1 comments - [hackernews](https://news.ycombinator.com/item?id=49776729) · 631 points · 707 comments ## Similar posts on daily.dev - [How ChatGPT serves ads. Here's the full attribution loop.](https://daily.dev/posts/how-chatgpt-serves-ads-here-s-the-full-attribution-loop--n31nsws5c) · Hacker News · 0 upvotes · 0 comments - [I didn't realize ChatGPT was serving ads until I dug into how it actually works](https://daily.dev/posts/i-didn-t-realize-chatgpt-was-serving-ads-until-i-dug-into-how-it-actually-works-iywrshaes) · XDA Developers · 2 upvotes · 0 comments --- Tags: [#openai](https://daily.dev/tags/openai), [#privacy](https://daily.dev/tags/privacy), [#chatgpt](https://daily.dev/tags/chatgpt) [View this post on daily.dev](https://daily.dev/posts/chatgpt-now-knows-what-you-do-on-other-websites-via-ad-collector-6ydoidhyl) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"ChatGPT now knows what you do on other websites via ad collector","url":"https://daily.dev/posts/chatgpt-now-knows-what-you-do-on-other-websites-via-ad-collector-6ydoidhyl","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/chatgpt-now-knows-what-you-do-on-other-websites-via-ad-collector-6ydoidhyl"},"datePublished":"2026-09-21T03:02:19.898Z","dateModified":"2026-09-21T05:48:20.966Z","description":"OpenAI's ad measurement system sets a cookie called __obi, scoped to .openai.com with SameSite=None, that gets sent back to OpenAI whenever a user visits any...","image":"https://media.daily.dev/image/upload/s--0_ODbtD2--/f_auto/v1722860399/public/Placeholder%2008","thumbnailUrl":"https://media.daily.dev/image/upload/s--0_ODbtD2--/f_auto/v1722860399/public/Placeholder%2008","isAccessibleForFree":true,"articleSection":"Lobsters","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Lobsters","logo":"https://media.daily.dev/image/upload/s--tl8v_Fku--/f_auto,t_logo/v1698841318/logos/lobste.jpg","url":"https://daily.dev/sources/lobsters"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/chatgpt-now-knows-what-you-do-on-other-websites-via-ad-collector-6ydoidhyl","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"openai,privacy,chatgpt","timeRequired":"PT7M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Lobsters","item":"https://daily.dev/sources/lobsters"},{"@type":"ListItem","position":3,"name":"ChatGPT now knows what you do on other websites via ad collector"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/chatgpt-now-knows-what-you-do-on-other-websites-via-ad-collector-6ydoidhyl#faq","mainEntity":[{"@type":"Question","name":"What is the __obi cookie that OpenAI sets and why does it get sent to other websites?","acceptedAnswer":{"@type":"Answer","text":"__obi is a cookie set by OpenAI's ad collector at bzr.openai.com, scoped to .openai.com with SameSite=None and Secure, which allows it to be attached to cross-site requests. It is created from a signed JWT issued by ChatGPT's backend that binds the identifier to a ChatGPT account or anonymous session, then gets sent to OpenAI whenever a page running OpenAI's advertiser conversion pixel loads, letting OpenAI link browsing on that site to a ChatGPT identity. Anyone auditing third-party ad pixels for cross-site identity leaks can track findings like this on daily.dev."}},{"@type":"Question","name":"Does OpenAI's ad tracking cookie __obi still work if a ChatGPT user is not logged in?","acceptedAnswer":{"@type":"Answer","text":"Yes, the mechanism works for logged-out users. Across 932 decoded sync tokens, 736 carried subject_type account_user and 196 carried anonymous, and the anonymous subject proved just as stable as the account-linked one, persisting on a device for at least 27 days, meaning cross-site tracking continues even without an active login. Developers weighing consent and identifier persistence in their own products can follow analysis like this on daily.dev."}},{"@type":"Question","name":"Does OpenAI's advertiser pixel collect personal data like email and phone number from advertiser websites?","acceptedAnswer":{"@type":"Answer","text":"Yes, the pixel SDK scrapes identity data directly from advertiser pages, not just from data advertisers deliberately supply. It hijacks the tag-manager data layer (window.dataLayer, adobeDataLayer, renamed GTM layers) to pull emails and phone numbers, and in observed traffic scraped identity events outnumbered advertiser-supplied ones 685 to 255; email, phone, and names are SHA-256 hashed, while country, region, city and postal code are sent unhashed. Teams reviewing what their ad-pixel integrations actually collect can keep up with findings like this on daily.dev."}}]} ```