404 Media Tests Show How ChatGPT Ads Target Users by Prompts, Emails and Phone Numbers
404 Media ran its own ChatGPT ads and found advertisers can steer them with conversation “context hints” and uploaded email, phone or Android ad-ID lists. Per OpenAI, inclusion targeting needs at least 25,000 matched users, and turning off ads personalization still allows ads based on the current chat.
By 1EyeAI Desk ·
404 Media reported on Oct. 15, 2026 how ChatGPT ad targeting works, based on two ads the outlet ran itself. When Free or Go users enter a prompt, ChatGPT often shows a related ad below the answer; 404 Media notes OpenAI began showing ads alongside image-generation results this month.
When setting up a campaign, advertisers can give OpenAI “context hints” describing relevant conversations or topics. OpenAI’s ad site says these hints “guide matching but aren't exact-match targeting rules.” 404 Media’s test ads targeted privacy and AI topics and linked to its earlier reporting on contractors reading ChatGPT chats.
Advertisers can also upload custom audiences. Per OpenAI’s help page, lists can contain emails, phone numbers, SHA-256 hashes of either, or Google Advertising IDs (GAID). Inclusion targeting and audience bid adjustments require at least 25,000 matched users; smaller audiences can be used only for exclusions, and advertisers can’t see which individuals matched. OpenAI says uploaded files are deleted after processing, usually within 24 hours. Asked whether it hashes raw uploads, as Meta does, OpenAI did not directly answer, per 404 Media.
Users can turn off ads personalization, but the setting itself says ads “may still be based on your current chat.” A third-party site, ChatGPT Ad Library, logs which prompts trigger which ads; it says its figures reflect its own captured sample, not OpenAI data.
What this means for you: On free ChatGPT tiers, what you type can shape the ads you see, and an advertiser holding your email or phone number may be able to target you.