OpenAI has begun testing what it calls negative targeting guidance in ChatGPT ads with a handful of advertisers, giving them a way to name the kinds of conversations they don’t want to run alongside, according to Trishla Ostwal at ADWEEK, which calls the feature exclusion targeting. Spokesperson Taya Christianson confirmed the test but wouldn’t give a timeline. Buyers have complained they can’t describe their audience to ChatGPT, can’t steer placement, and can’t see where their ads ended up, even at CPMs that Seer Interactive has tracked as high as $60 on B2B campaigns. An SE Ranking study in July across more than 50,000 queries found ads on 26% of commercial prompts, one in seven of which had nothing to do with the question, rising to half in news, politics, and relationship chats.
