ChatGPT ads skew toward lower-income users, study of 91 accounts finds

The research, titled "The Beginning of ChatGPT Ads," offers an early look at how OpenAI's ad system behaves in practice. By simulating users across income and racial lines and tracking daily interactions from February 2026, the team captured a pattern: ad delivery ramped up roughly two weeks after account creation and skewed toward lower-income profiles. That contrasts with the absence of any racial skew, suggesting income — not race — is the variable shaping exposure.

The content mix also reveals a contextual approach, with more than half of ads tied to retail and IT services and food-delivery promotions appearing alongside recipe questions. Meanwhile, sensitive categories like medical, mental health and political topics stayed largely ad-free, consistent with OpenAI's stated rules. For businesses buying ChatGPT ad inventory, the findings hint that reach may concentrate among certain income brackets, which could affect campaign planning and audience assumptions.