404 Media Tests: Advertisers Can Target ChatGPT Users by Email Address and Phone Number

What happened

404 Media ran two paid test ads inside ChatGPT's ad system and confirmed advertisers can upload “raw” email addresses and phone numbers to build targeting audiences, with OpenAI checking whether the data matches real ChatGPT users; “hashed” (digitally obscured rather than raw) emails and phone numbers, as well as Android ad IDs, work too. Targeting requires at least 25,000 matched users; smaller lists can only be used to exclude audiences.

Another targeting route is “context hints”: advertisers describe conversation topics or keywords in text, which OpenAI's documentation says “guide matching but aren't exact-match targeting rules.” OpenAI told 404 Media that uploaded files are typically deleted within 24 hours and that advertisers cannot see exactly which individuals were matched.

ChatGPT's settings offer an opt-out for “ads personalization,” but the settings text explicitly states that “ads may still be based on your current chat.” Background: ChatGPT ads launched in the US in February 2026, and its ad library snapshot reportedly already holds at least 43,410 ads and 415,289 placements.

“Ads may still be based on your current chat.”

— ChatGPT settings text (as reported by 404 Media / MobileSyrup)

Key facts

1. Contact targeting: Upload raw emails/phones, hashed contact details or Android ad IDs for targeting
2. 25,000-user threshold: Targeting requires at least 25,000 matched users; smaller lists can only exclude audiences
3. Context hints: Advertisers describe chat topics/keywords in text — “guide matching but aren't exact-match targeting rules”
4. 24-hour deletion: OpenAI says uploaded files are typically deleted within 24 hours; advertisers can't see who was matched
5. Still served when off: With “ads personalization” off, ads may still be based on the current chat