OpenAI Expands ChatGPT Ad Pilot With Criteo Partnership

OpenAI is integrating advertisements into the free and Go tiers of ChatGPT, marking a shift toward a model that CEO Sam Altman once described as a "last resort." The company has named Criteo as its first official ad tech partner to help power these placements, which are currently appearing in roughly one out of every five conversation threads in the U.S. [1, 2]

The ads are tailored to the specific topic of a user's prompt and appear as buttons at the bottom of the chatbot’s output. For instance, questions regarding the gig economy have triggered ads for Uber, while travel-related queries have surfaced links to Booking.com [1]. OpenAI claims these ads do not influence the content of the AI’s responses and that user conversation data is not shared with advertisers. However, the company notes that placements are influenced by a user’s prompt history and the information stored in the chatbot’s memory [1].

For Criteo, the partnership serves as a validation of its strategy to become the commerce intelligence layer for generative AI. By integrating its commerce data and Model Context Protocol server, Criteo aims to provide the transaction-linked signals necessary for high-quality product recommendations within conversational interfaces [2]. This allows marketers to tap into Criteo’s audience tools directly through AI agents, effectively moving the front end of advertising into the chatbot experience [2].

OpenAI’s move into advertising comes as the company seeks to monetize its user base while managing the high costs of supporting free users. While the company reports that it has seen no negative impact on consumer trust metrics and low dismissal rates for the ads, the transition carries significant risks [1]. Marketing experts warn that if the ads degrade the quality of the user experience or undermine the perceived neutrality of the chatbot, OpenAI could drive users toward competitors like Google’s Gemini or Anthropic’s Claude, neither of which currently feature sponsored buttons in their outputs [1].

As OpenAI prepares to expand this pilot to Canada, Australia, and New Zealand, the company is actively hiring for roles dedicated to mitigating the risks of this integration, including product safety and user trust [1]. The central question remains whether OpenAI can successfully scale its ad business without eroding the personal, private environment that has defined the ChatGPT experience for its millions of users.