Signals/ ChatGPT & OpenAI/ Brief /
ChatGPT ads can be aimed at you by email, phone number or your prompts
A 404 Media test shows advertisers can reach specific people inside ChatGPT, and the ad system can also use what you type.
On October 15, 404 Media published its own test of the ChatGPT advertising system. Its finding: advertisers can pick groups of ChatGPT users by email address, phone number or a unique ID tied to a phone. The report also says targeting can draw on what users type into prompts. ChatGPT has about one billion monthly users, per the same article.
The email and phone option is not brand new. Search Engine Roundtable reported in July that OpenAI added an “Audiences” feature, where advertisers upload raw or hashed emails and phone numbers to use as campaign filters. Many ad platforms already work this way.
The catch: The visible part of the 404 Media article included no OpenAI response and no opt-out details, and part of it is paywalled. I could not confirm what controls OpenAI offers, so check the official settings yourself.
What to do now
- Open ChatGPT’s settings and look for privacy or personalization options. Menu names may differ, so check what your app shows.
- Avoid putting health, money or legal details into a chat you would not want used for matching.
- If you run a small business, read OpenAI’s ad policies before uploading customer lists, and use only contacts who agreed to marketing.
