Can advertisers target you in ChatGPT by email or phone number?
The short answer
Email and phone lists: confirmed in OpenAI's own documentation. Advertisers can upload emails, phone numbers, SHA-256 hashes of either, or Android advertising IDs OpenAI Help, Custom audiences (checked 15 Oct 2026). 404 Media tested the system and reported it on 15 October 2026 404 Media (15 Oct 2026).
The limit: a list needs at least 25,000 matched users before it can be used to show ads. Smaller lists can only be used to exclude people OpenAI Help, Custom audiences (checked 15 Oct 2026).
Your chats: ads are picked mainly from the current conversation. With ad personalization on, past chats and memory can count too OpenAI Help, Ads in ChatGPT (checked 15 Oct 2026).
Not found: we found no age, gender or interest targeting in OpenAI's campaign documentation OpenAI Help, Create campaigns (checked 15 Oct 2026). OpenAI publishes no ad library.
Yes, as part of a large list. OpenAI calls the feature custom audiences (lists of customers or prospects that an advertiser uploads). It accepts five identifier types OpenAI Help, Custom audiences (checked 15 Oct 2026). They are email address, phone number, a SHA-256 hash of either, and Google Advertising ID (GAID, a resettable ID on Android phones). An advertiser can include a list, exclude it, or bid more for people on it OpenAI Help, Custom audiences (checked 15 Oct 2026).
What OpenAI says protects you. "Ads Manager does not show individual matched users or let you select specific people from an audience" OpenAI Help, Custom audiences (checked 15 Oct 2026). "Uploaded audience files are deleted after processing, usually no later than 24 hours after upload" OpenAI Help, Custom audiences (checked 15 Oct 2026). OpenAI also told 404 Media: "Being on an advertiser's list does not, by itself, cause an ad to be shown." 404 Media (15 Oct 2026). These are OpenAI's own statements; no outside audit has been published.
What is still unanswered. OpenAI accepts raw, unhashed emails and phone numbers. 404 Media asked whether OpenAI hashes them on arrival; OpenAI did not answer 404 Media (15 Oct 2026).
The headline you may have seen, "ChatGPT Can Target Ads With Your Prompts, Email, and Phone", comes from Gadget Review's write-up of the 404 Media report Gadget Review (15 Oct 2026).
Who sees ChatGPT ads, and since when?
Free and Go users, and some logged-out visitors. Plus, Pro, Business, Enterprise and Edu accounts get no ads OpenAI Help, Ads in ChatGPT (checked 15 Oct 2026). Logged-out visitors can see ads; OpenAI says these are "appropriate for all audiences" OpenAI Help, Ads in ChatGPT (checked 15 Oct 2026). That is wider than the first test, which was for "logged-in adult users on the Free and Go subscription tiers" OpenAI (9 Feb 2026, updated 11 Aug 2026).
Not shown: to accounts OpenAI identifies as under 18 OpenAI Help, Ads in ChatGPT (checked 15 Oct 2026), and in Temporary Chats. OpenAI's help page says: "Temporary Chats will not show ads." OpenAI Help, Ads in ChatGPT (checked 15 Oct 2026)
Format. Ads sit below a ChatGPT answer, labelled as sponsored and set apart from it OpenAI Help, Ads in ChatGPT (checked 15 Oct 2026). Each shows an advertiser name, icon, title, text, image and link OpenAI Help, Ads basics (checked 15 Oct 2026). 404 Media reported that OpenAI recently added ads next to image-generation results 404 Media (15 Oct 2026). If you message an advertiser through an ad, that advertiser sees those messages OpenAI Help, Ads in ChatGPT (checked 15 Oct 2026).
How did the rollout go?
- 9 Feb 2026: test starts in the US OpenAI (9 Feb 2026, updated 11 Aug 2026) Axios (9 Feb 2026).
- 3 Mar 2026: Criteo confirmed as the first ad-tech partner in the US pilot Digiday (3 Mar 2026).
- 26 Mar 2026: OpenAI says expansion will start with Canada, Australia and New Zealand OpenAI (9 Feb 2026, updated 11 Aug 2026).
- May 2026: self-serve Ads Manager opens to US businesses on 5 May; custom audiences appear in a gated rollout on 14 May, per trade press PPC Land (14 Jul 2026).
- 10 Jun 2026: LiveRamp partnership to measure in-store purchases Adweek (10 Jun 2026).
- 11 Aug 2026: ads live in the UK, Mexico, Brazil, Japan and South Korea OpenAI (9 Feb 2026, updated 11 Aug 2026).
- 18 Aug 2026: OpenAI announces ads in 31 European markets "next week" OpenAI (18 Aug 2026).
- 15 Oct 2026: OpenAI's advertiser availability page lists 56 countries for self-serve buying OpenAI Help, Ads availability (checked 15 Oct 2026). That is where advertisers can sign up; it does not prove ads are showing in each.
What targeting options do ChatGPT advertisers get? (at a glance)
ChatGPT ad targeting options in OpenAI's documentation, as of 15 October 2026.
| Option | What OpenAI documents | Status |
|---|---|---|
| Conversation context | The current chat's topic and intent, matched to the ad and its landing page OpenAI Help, Ads in ChatGPT (checked 15 Oct 2026) | Always on |
| Context hints | Topics an advertiser describes; "not exact-match keywords" OpenAI Help, Ads basics (checked 15 Oct 2026) | Available |
| Past chats, memory, ad history | Used only with ad personalization on OpenAI Help, Ads in ChatGPT (checked 15 Oct 2026) | You can switch off |
| Location | Country, plus states, cities, markets or postal codes where available OpenAI Help, Create campaigns (checked 15 Oct 2026) | Available |
| Platform | Android app, Android web, desktop web, iOS app, iOS web OpenAI Help, Create campaigns (checked 15 Oct 2026) | Available |
| Custom audiencesemail, phone, hashes, GAID | Include, exclude or bid up; 25,000 matched users to include OpenAI Help, Custom audiences (checked 15 Oct 2026) | Not for EEA or Swiss campaigns OpenAI developer docs (checked 15 Oct 2026) |
| Age, gender, interests | Not listed in campaign settings OpenAI Help, Create campaigns (checked 15 Oct 2026) | None found |
| Under-18 accounts | No ads OpenAI Help, Ads in ChatGPT (checked 15 Oct 2026) | Excluded |
| Political ads | "we currently do not allow political advertising in ChatGPT." OpenAI Help, Ads in ChatGPT (checked 15 Oct 2026) | Not allowed |
Reading the table. These are the options OpenAI documents for advertisers. Absence from the documentation does not prove a signal is never used. Health and finance advertisers can be approved case by case in the US OpenAI Ad Policies (10 Sep 2026).
How does ChatGPT pick an ad from your conversation?
Your current chat comes first. OpenAI weighs the chat's context and intent, plus the ad's text and landing page OpenAI Help, Ads in ChatGPT (checked 15 Oct 2026). It also uses the advertiser's context hints and targeting, and general location and language OpenAI Help, Ads in ChatGPT (checked 15 Oct 2026). Ads then compete in a "relevance-weighted, second-price auction" OpenAI Help, Ads basics (checked 15 Oct 2026).
Personalization adds your history. With ad personalization on, OpenAI can use select signals from your wider ChatGPT use. "If memory is on, ChatGPT may save and use memories and reference recent chats when selecting an ad." OpenAI Help, Ads in ChatGPT (checked 15 Oct 2026). The privacy policy adds: "For Free and Go users, we also collect information about your ads history and interests" OpenAI Privacy Policy (10 Sep 2026).
Personalization off is not ads off. OpenAI says that with it off, "we won't use other chat threads, ads history or topics to inform the ads we show you." OpenAI Help, Ads in ChatGPT (checked 15 Oct 2026). The setting itself warns: "Ads may still be based on your current chat" 404 Media (15 Oct 2026). OpenAI's pages do not say whether this setting also removes you from advertiser lists.
Sensitive conversations are excluded. "Ads are not eligible to appear near sensitive topics, including personal health, mental health, or politics." OpenAI Help, Ads in ChatGPT (checked 15 Oct 2026). The ad policies also list emotionally reliant chats and topics such as self-harm, hate and weapons as off-limits OpenAI Ad Policies (10 Sep 2026).
Data can flow in from advertisers. OpenAI's privacy policy says it receives information from advertisers and data partners. "For example, we could receive information about purchases you make from these advertisers." OpenAI Privacy Policy (10 Sep 2026). The LiveRamp deal is for measuring in-store purchases after an ad Adweek (10 Jun 2026).
How does email and phone list matching work?
The same method across big ad platforms. The advertiser cleans each email or phone number into one standard form. It then uploads it, often as a SHA-256 hash (a fixed 64-character fingerprint of the text). The platform makes the same fingerprint from its own users' details and looks for equal pairs. Google describes exactly this for Customer Match Google Ads Help (checked 15 Oct 2026).
OpenAI's rules for the clean-up. Emails are trimmed and lowercased. Phone numbers must be in E.164 form (a plus sign, country code and number, no spaces) OpenAI Help, Custom audiences (checked 15 Oct 2026). A processed list usually takes 20 to 30 minutes OpenAI Help, Custom audiences (checked 15 Oct 2026).
Customer-list targeting compared. A dash means we could not confirm it from a page we could read.
| Point | ChatGPT custom audiences | Google Customer Match | Meta Custom Audiences |
|---|---|---|---|
| Identifiers | Email, phone, SHA-256 of either, GAID OpenAI Help, Custom audiences (checked 15 Oct 2026) | Email, phone, name and address Google Ads Help (checked 15 Oct 2026) | Personal data such as emails and names Meta developer docs (checked 15 Oct 2026) |
| Hashing | Raw or SHA-256 accepted; whether raw uploads are hashed: not answered 404 Media (15 Oct 2026) | Advertiser or Google hashes with SHA-256 Google Ads Help (checked 15 Oct 2026) | "You must hash data as SHA256" Meta developer docs (checked 15 Oct 2026) |
| Size to target | 25,000 matched users; none to exclude OpenAI developer docs (checked 15 Oct 2026) | At least 100 members in the last 540 days; no "overly narrow" lists Google Ads policy (checked 15 Oct 2026) | — |
| Data source rule | First-party only; "Don't upload broker-sourced data" OpenAI developer docs (checked 15 Oct 2026) | Collected "in the first-party context" Google Ads policy (checked 15 Oct 2026) | — |
| Upload kept | Usually deleted within 24 hours OpenAI Help, Custom audiences (checked 15 Oct 2026) | Marked for deletion after matching and checks Google Ads Help (checked 15 Oct 2026) | — |
| Who matched | Not shown; counts only as ranges OpenAI developer docs (checked 15 Oct 2026) | — | — |
What this means in practice. OpenAI's 25,000 floor is far higher than the 100 members Google's policy states Google Ads policy (checked 15 Oct 2026). Exclusion lists have no size floor at OpenAI OpenAI developer docs (checked 15 Oct 2026). OpenAI's developer docs say advertisers see matched counts only as ranges, not exact numbers OpenAI developer docs (checked 15 Oct 2026).
Why is a hashed email not anonymous?
A hash is a stable stand-in for you. The US Federal Trade Commission's technologists put it plainly: "the same input data will always create the same hash." FTC Office of Technology (24 Jul 2024). So the hash of your email works as a lasting ID that two companies can match.
For example, the SHA-256 hash of [email protected] is always:
86e0b9e56c17cc4d12387e1949b85053fbe73bc3ce5a1188713a9d300cc6133d
Anyone holding a list of emails can compute the same string and find the match. Phone numbers are worse. There are few enough possible numbers that the FTC notes they can be guessed and checked until the original turns up FTC Office of Technology (24 Jul 2024).
The FTC's test: "data is only anonymous when it can never be associated back to a person." FTC Office of Technology (24 Jul 2024). Hashing helps keep raw emails out of transit. It does not stop matching, which is its whole purpose in advertising.
What are the privacy risks for ChatGPT users?
1. Lists you never agreed to. OpenAI forbids broker data: Audience Data must be first-party and not "purchased, licensed" OpenAI Ad Tools Terms (24 Aug 2026). The advertiser must also hold the required consents OpenAI Ad Tools Terms (24 Aug 2026). These are promises the advertiser makes. We found no description of how OpenAI checks them.
2. Old sign-ups follow you. Any company you gave your email or phone number to can upload it, if it holds the consents the terms require OpenAI Ad Tools Terms (24 Aug 2026). If your ChatGPT login uses that same email, a match is possible.
3. Inferences from your chats. Ads are chosen from what you talk about, and with personalization on, from past chats and memory OpenAI Help, Ads in ChatGPT (checked 15 Oct 2026). Advertisers never see the chat, OpenAI says OpenAI Help, Ads in ChatGPT (checked 15 Oct 2026). But clicking an ad on a topic tells that advertiser you were interested in it.
4. Data coming back in. OpenAI may receive purchase data from advertisers OpenAI Privacy Policy (10 Sep 2026), and partners like LiveRamp connect ads to store sales Adweek (10 Jun 2026).
The critics' view. Zach Edwards of DecryptAds told 404 Media that "they are boiling the frog and hoping their users won't notice the privacy implications from their new ads" 404 Media (15 Oct 2026). He called for stronger data-broker laws covering advertisers that share personal data 404 Media (15 Oct 2026).
How do you limit ad targeting in ChatGPT?
- Open the controls. "Ad controls are available in Settings > Ad Controls" OpenAI Help, Ads in ChatGPT (checked 15 Oct 2026). From there you can turn off personalization, hide ads, see why an ad appeared, and clear ads data OpenAI Help, Ads in ChatGPT (checked 15 Oct 2026).
- Turn off ad personalization. Past chats, ad history and topics then stop counting; the current chat still does OpenAI Help, Ads in ChatGPT (checked 15 Oct 2026).
- Review memory. Memory can feed ad choice when it is on OpenAI Help, Ads in ChatGPT (checked 15 Oct 2026). Our AI privacy guide covers memory and training settings.
- Delete ads data. OpenAI says deleted ads data "is retained for up to 30 days before being removed from our servers." OpenAI Help, Ads in ChatGPT (checked 15 Oct 2026). Your chats are not affected OpenAI Help, Ads in ChatGPT (checked 15 Oct 2026).
- Use Temporary Chats for sensitive questions. They show no ads OpenAI Help, Ads in ChatGPT (checked 15 Oct 2026).
- Opt out of OpenAI's marketing sharing. The privacy policy offers a marketing privacy control, and honours Global Privacy Control (a browser opt-out signal) OpenAI Privacy Policy (10 Sep 2026).
- Use a separate email for ChatGPT. This is our suggestion, not OpenAI's. List matching needs your ChatGPT login to equal what an advertiser holds. A dedicated or masked address makes that much less likely.
- Shrink the lists you are on. Remove yourself from people-search sites, and if you live in California use the DROP broker-deletion system.
What can investigators learn from ChatGPT ads?
Less than from other ad platforms, for now. OpenAI does not publish a searchable ad archive. Third-party trackers fill some of the gap. LLM Pulse, a vendor, says its library records ads seen in its own prompt runs; it calls the result a sample, not a census LLM Pulse (vendor, 28 Sep 2026). Another site, ChatGPT Ad Library, told 404 Media it had logged at least 43,410 ads across 415,289 placements (its own figures) 404 Media (15 Oct 2026).
An EU ad repository is due. The European Commission designated ChatGPT a Very Large Online Search Engine on 31 August 2026. It must meet the extra duties "i.e. by January 2027" European Commission (31 Aug 2026). Article 39 of the Digital Services Act requires such services to publish an ad repository DSA Article 39 text. That repository must include the main targeting criteria and total reach DSA Article 39 text.
What a single ad tells you. On your own account, "About this ad" in the ad's menu explains why it appeared OpenAI Help, Ads in ChatGPT (checked 15 Oct 2026). Ad controls also list the topics used for relevance OpenAI Help, Ads in ChatGPT (checked 15 Oct 2026). We could not confirm whether it says when an advertiser's list was involved.
What it does not tell you. Seeing an ad does not prove you are on a list, because list membership alone does not trigger an ad, OpenAI says 404 Media (15 Oct 2026). Advertisers see only grouped figures such as views and clicks OpenAI Help, Ads in ChatGPT (checked 15 Oct 2026). If the system works as OpenAI describes, an advertiser cannot confirm that a given person uses ChatGPT.
Useful questions for research. Which brands advertise against which topics? Which sensitive topics do ads skip? Record the prompt, date, country and screenshot each time. Our OSINT OPSEC guide covers running research accounts safely.
Can you use ad targeting to find or watch a specific person?
No, and the platforms forbid it. OpenAI's Ad Tools Terms bar using its tools or data to "identify or re-identify any OpenAI user" or to "infer sensitive or other information about any person" OpenAI Ad Tools Terms (24 Aug 2026). Uploaded data must be first-party with proper consent OpenAI Ad Tools Terms (24 Aug 2026). Google bars Customer Match lists aimed at an "overly narrow or specific audience" Google Ads policy (checked 15 Oct 2026).
The design works against it too. OpenAI requires 25,000 matched users to target and shows no individual matches OpenAI Help, Custom audiences (checked 15 Oct 2026). Uploading someone's email to probe for them would breach the terms. It may also breach privacy law where you are.
The investigator's line. Use ads as a public record of what is advertised, to whom, and around which topics. Do not use ad systems to locate, track or profile one person. Our guide Is OSINT legal? sets out the wider limits. This page is general information, not legal advice.
Sources
Every source was read on 15 October 2026. OpenAI, Google and Meta pages describe their own products. This page is general information, not legal advice.
- OpenAI Help Center — "Ads in ChatGPT" (read 15 Oct 2026; page shows "Updated: yesterday")
- OpenAI Help Center — "Ads in ChatGPT: The Basics" (read 15 Oct 2026)
- OpenAI Help Center — "Set up custom audiences for your campaign" (read 15 Oct 2026)
- OpenAI developer docs — "Custom Audiences" (read 15 Oct 2026; undated)
- OpenAI Help Center — "Create campaigns for ChatGPT" (read 15 Oct 2026)
- OpenAI Help Center — "ChatGPT Ads availability" (read 15 Oct 2026)
- OpenAI — Ad Tools Terms (published 24 Aug 2026)
- OpenAI — Ad Policies (updated 10 Sep 2026)
- OpenAI — Privacy Policy (updated 10 Sep 2026)
- OpenAI — "Testing ads in ChatGPT" (9 Feb 2026, updates 26 Mar, 7 May and 11 Aug 2026)
- OpenAI — "ChatGPT Ads expands across Europe" (18 Aug 2026, update 31 Aug 2026)
- Joseph Cox, 404 Media — "How ChatGPT Targets Ads To You Based On Your Prompts, Phone Number, and Email" (15 Oct 2026)
- Madison Mills, Axios — "OpenAI tests ads in ChatGPT" (9 Feb 2026)
- Ronan Shields, Digiday — "Ad Tech Briefing: Criteo named first ad tech partner to OpenAI's ChatGPT ad pilot" (3 Mar 2026)
- Ryan Joe, Adweek — OpenAI partners with LiveRamp (10 Jun 2026; partly paywalled)
- Luís Rijo, PPC Land — "OpenAI blocks ChatGPT ad audiences under 25,000 matched users" (14 Jul 2026)
- Aminu Abdullahi, TechRepublic — "OpenAI Brings ChatGPT Ads to 31 European Countries" (20 Aug 2026)
- Alex Barrientos, Gadget Review — "ChatGPT Can Target Ads With Your Prompts, Email, and Phone" (15 Oct 2026)
- European Commission — designation of ChatGPT, Reddit and Roblox under the Digital Services Act (31 Aug 2026)
- Digital Services Act, Article 39 — additional online advertising transparency (text read 15 Oct 2026)
- Daniel Peris, LLM Pulse — "Introducing the World's First ChatGPT Ads Library" (vendor; 21 Aug 2026, updated 28 Sep 2026)
- FTC Office of Technology — "No, hashing still doesn't make your data anonymous" (24 Jul 2024)
- Google Ads — Customer Match policy (read 15 Oct 2026)
- Google Ads Help — "About the customer matching process" (read 15 Oct 2026)
- Meta for Developers — Marketing API, Custom Audiences guide (read 15 Oct 2026)
Frequently Asked Questions
Do ads change the answers ChatGPT gives?
OpenAI says no. Its February 2026 announcement states that ads do not influence the answers ChatGPT gives you. It adds that ads are always labelled as sponsored and shown apart from the answer. No independent test of that claim has been published that we could find.
Can I use ChatGPT for free without ads?
Yes. Free users can switch to an ad-free version from the ad controls. OpenAI says this removes ads but comes with lower usage limits and less access to some tools, such as image generation and deep research. Plus and Pro plans are ad-free.
Does clicking or reporting an ad get saved to my ChatGPT memory?
OpenAI says hiding, reporting, viewing or clicking an ad does not add to your memory. If you ask ChatGPT about an ad you saw, that chat can be referenced later when chat history reference is turned on.
Are ChatGPT ads personalized in Europe?
Not at first. OpenAI says personalized ads are not initially available in the European Economic Area or Switzerland. Its advertiser docs say custom audiences are not supported for campaigns targeting those countries. Ads in Europe began in late August 2026 for Free and Go users.
Can an advertiser get my email address from a ChatGPT ad?
Not from OpenAI, according to its help center. Advertisers do not receive your name, email, chats, memories, precise location or IP address. They get only grouped figures such as views and clicks. If you visit the advertiser's site, message it, or buy something, it gets whatever you give it there.
Does ChatGPT show ads to teenagers?
OpenAI says it does not show ads to accounts it identifies as belonging to users under 18. It decides this from account information and, where available, age prediction. Logged-out visitors may see ads that OpenAI says are suitable for all audiences.
Next step
Check your own ad settings now. In ChatGPT, open Settings, then Ad Controls. Turn off personalization if you want, clear your ads data, and look at the topics listed there. Then see how your email shows up in public with our email OSINT tools.