ChatGPT ad targeting is broader than most people assumed. According to a report from Gizmodo, based on testing by 404 Media, advertisers can aim ads at you using your prompts, and also using your email address or phone number.
The catch is the opt-out. OpenAI’s personalization switch doesn’t stop ads that are based on the chat you’re having right now.
Here are the key facts as reported on October 16, 2026.
- Advertisers can upload lists of emails, phone numbers or Android advertising IDs.
- OpenAI checks whether those match ChatGPT accounts, and matched users become eligible for an ad.
- A list must hold at least 25,000 entries, and OpenAI says it deletes uploaded files within 24 hours.
- An ads personalization setting exists, but it doesn’t cover your current chat.
The takeaway: switching off personalization helps, but it isn’t a full privacy shield.
How ChatGPT ad targeting actually works
There are two separate systems in play. One looks at what you’re talking about. The other starts from a list of people an advertiser already knows.
The first is context. Per MobileSyrup’s summary of the 404 Media test, advertisers can choose topics or keywords. The test targeted people asking about privacy or AI.
The second is the contact list. An advertiser uploads emails, phone numbers or device IDs, and OpenAI works out which ones belong to ChatGPT users. The advertiser reportedly isn’t told when a match happens.
| Targeting method | How it works | What OpenAI says |
|---|---|---|
| Current chat context | Advertisers pick topics or keywords that match what you are asking about. | Ads may still be based on your current chat, even with personalization off. |
| Past chats and activity | Ads are picked using your earlier chats, activity and preferences. | You can switch this off in your ads personalization setting. |
| Uploaded contact lists | Advertisers upload emails, phone numbers or Android ad IDs. OpenAI checks for matching accounts. | Uploaded files are deleted within 24 hours. Lists need at least 25,000 entries. |
That table is the part to remember. Only one of the three rows has a clear off switch.
Ads are also no longer just text. Gizmodo notes OpenAI recently added visual promos next to the text ads inside conversations, so the placements are getting more prominent, not less.
Does the ChatGPT ads opt-out really work?
Partly. OpenAI says you can turn off personalized ads in settings, and that this stops ads based on your past chats, activity and preferences.
But OpenAI also says ads “may still be based on your current chat.” So if you ask about, say, a new laptop, a laptop ad can still appear.
Gizmodo also notes that the setting seems to cover chat history and activity, not uploaded identifiers. In other words, being on an advertiser’s list may be a separate matter. OpenAI also says being on a list doesn’t guarantee you’ll see an ad.
What your phone number and email have to do with it
If you signed up for ChatGPT with an email or phone number, that identifier is exactly what a matching system can look for. You didn’t hand it to an advertiser. They brought it from somewhere else.
That could be a store loyalty list, a newsletter or a data broker. The ad shows up inside a tool you use for private questions, which is why this feels different from a banner on a news site.
Third-party tools, such as a ChatGPT ad library site mentioned in the MobileSyrup piece, reportedly make setting up these campaigns easier. That lowers the bar for smaller advertisers too.
It’s also unclear whether OpenAI hashes raw uploads before matching, according to Gizmodo. That’s a question OpenAI hasn’t answered publicly yet, as far as the report shows.
Practical ways to limit ChatGPT ad targeting
You can’t fully block this, but a few habits shrink what advertisers can use. Try these in order.
- Open ChatGPT settings and turn off the ads personalization option.
- Use a separate email address for AI accounts, not the one on your shopping and loyalty cards.
- Keep sensitive topics, like health or money worries, out of chats that may carry ads.
- Clear or turn off chat history if you don’t need it.
Why this matters: each step cuts a different link between you and an advertiser’s list.
If you already use ChatGPT for money questions, our guide to Finances in ChatGPT covers what that feature sees. For the earlier ad rollout, see how image generation ads work and which setting to change.
Frequently Asked Questions
Can advertisers see my ChatGPT conversations?
The reports don’t say advertisers read your chats. They describe topic and keyword matching, so ads can follow what you discuss without the advertiser seeing the conversation itself.
Will turning off ad personalization stop all ads?
No. It limits ads based on past chats and activity. Ads can still appear, and they may still use your current chat.
How many people does an advertiser need to target?
According to the reports, a custom audience needs a list of at least 25,000 entries. That makes pinpointing one person unlikely, though not impossible to reach as part of a larger list.
Does OpenAI keep the contact lists?
OpenAI says it deletes uploaded files within 24 hours of processing them. Whether raw data is hashed first is unclear.
My take
Treat any chatbot with ads like a search engine with ads: helpful, but not private. Turn off personalization today, keep a separate email for AI tools, and think twice before pasting anything sensitive. If OpenAI later explains how uploads are hashed and how the opt-out will cover them, that would be the real fix.
Leave a Reply