Your ChatGPT is quietly becoming an ad targetHow ChatGPT Targets Ads To You Based On Your Prompts, Phone Number, and Email

404 Media tested ChatGPT's ad system and confirmed advertisers can upload email addresses, phone numbers, or Android ad IDs to target users, and ads can also be keyed to the prompts they type.

404 Media ran two ads inside ChatGPT and found that advertisers can upload raw or hashed email addresses, phone numbers, Android advertising IDs, or the hashed versions of personal data. OpenAI matches these against real ChatGPT users, then targets those people with ads.

Uploaded lists must contain more than 25,000 entries before they can be used for inclusion targeting; otherwise users can only be marked for exclusion. OpenAI said files are deleted within 24 hours of processing and that advertisers cannot see which individuals matched.

Your prompts are being read by ad buyers

OpenAI also lets advertisers provide context hints describing the kinds of conversations their products might be relevant to. For its own campaign, 404 Media used the context: people discussing privacy, the privacy of ChatGPT chats, or currently asking about technology and AI topics.

This means everything you type into ChatGPT can be used to decide which ad you see. Previous reporting has already shown that human reviewers read real users' chats.

Big Tech is boiling the frog

Zach Edwards of DecryptAds said OpenAI's ad products are operating like most other big companies that start monetising their audiences: they are boiling the frog and hoping users won't notice the privacy implications of the new ads.

He argued for stronger data broker laws, so that advertisers sharing sensitive personally identifiable information with ad tech companies such as OpenAI would need to register as data brokers and offer people a clear way to remove their data.