ChatGPT has new ad formats. Heres what they look like.

OpenAI is introducing visual ads to its ChatGPT platform. The company announced this development in a blog post published on Monday. In addition to the new ad format, OpenAI also unveiled expanded measurement partnerships for advertisers and pilot programs to demonstrate how its ad placement safeguards operate.

According to OpenAI, ChatGPT currently reaches 1.2 billion people weekly. The company views advertising as a means to fund broader access to AI beyond what subscription fees can cover. The mockups provided by OpenAI showcase two ad layouts, both centered around a fictional brand named Heirloom Grocery.

In the first layout, a single banner with a headline sits atop a concise product description and a "Learn more" button. The second layout replaces the banner with a scrollable row of photos. In both cases, the ad bears an "Ad" label and appears beneath a progress indicator for an image ChatGPT is still generating. OpenAI plans to roll out the format by initially testing the ads during image generation, ensuring they remain clearly labeled and separate from the image being created.

OpenAI reiterated that advertising does not influence ChatGPT's responses. Advertisers can now utilize a feature called Negative Phrases to set their placement exclusions, thereby qualifying for the platform. OpenAI is also conducting brand suitability pilots with DoubleVerify and Integral Ad Science. These pilots enable external firms to evaluate OpenAI's safeguards in a controlled test environment without accessing real user conversations.

The company's placement guardrails are designed to keep ads out of emotionally vulnerable or sensitive conversations. To this end, qualifying advertisers can employ Negative Phrases to establish their own placement exclusions. Additionally, OpenAI is running brand suitability pilots in collaboration with DoubleVerify and Integral Ad Science. These pilots permit these external firms to assess OpenAI's safeguards in a controlled environment while refraining from accessing real user conversations.

Lidiane Jones, CEO of Integral Ad Science, emphasized the importance of brand safety protections in a statement accompanying the announcement. She highlighted the need for durable, consistent, and scalable brand safety measures. The testing phase, slated to commence later this month, will initially involve a select group of advertisers in the United States.