OpenAI to test labeled visual ads while ChatGPT users generate images
OpenAI says ChatGPT will show clearly labeled image ads during image generation, tested with US users and a first group of advertisers later this month. It says ads do not influence answers.
OpenAI said on 5 October that it will begin testing a visual ad format in ChatGPT later this month, with the ads appearing while a user's image is being generated. BleepingComputer, which reported the announcement, says the test starts with users in the United States and an initial group of advertisers.
According to BleepingComputer's account of the announcement, the ads will use images to show product inspiration, product use or service experiences. OpenAI says they will be "clearly labeled" and "remain separate from the image being created", and that "ads do not influence ChatGPT's answers". Those are OpenAI's assurances; no outside party has tested them.
The company is also adding measurement for advertisers, BleepingComputer reports, working with Hightouch, Tealium and LiveRamp on conversion tracking. It says it is working with DoubleVerify and Integral Ad Science on brand safety, to keep ads away from sensitive conversations. How those filters decide what counts as sensitive is not described in the reporting.
BleepingComputer puts ChatGPT's reach at 1.2 billion weekly users, a figure it attributes to the announcement; this article could not open OpenAI's own post, which the outlet links as the primary source. Pricing, the size of the test and the date it starts were not given in the account, and no other outlet had independently confirmed the details at the time of writing.