OpenAI has paused in-app "suggestions" within its ChatGPT application after users, particularly paying subscribers, reported seeing messages resembling advertisements for third-party companies. Mark Chen, OpenAI's Chief Research Officer, acknowledged that the company "fell short" with these promotional messages and stated they have been deactivated to improve precision and user experience.
Subscribers to ChatGPT's paid tiers publicly complained about seeing suggestions for entities such as Peloton and Target. These users expressed skepticism regarding the nature of these prompts, which appeared within their conversational AI interface. OpenAI initially stated that these were tests to display applications built on the ChatGPT app platform, which was announced in October, and maintained that there was "no financial component" to the suggestions.
The company's responses to the controversy have varied among its executives. Nick Turley, head of ChatGPT, publicly denied that the platform was running "live tests for ads," asserting that any circulating screenshots were either inaccurate or did not depict advertisements. However, Chen adopted a more conciliatory tone, acknowledging that the issue transcended mere user confusion. He specified that the suggestions have been turned off "while we improve the model's precision" and indicated that better user controls for these features are under consideration.
The incident occurs amidst broader strategic movements at OpenAI. Earlier this year, Fidji Simo, former Instacart and Facebook executive, joined OpenAI as CEO of Applications, with market observers anticipating her role in developing the company's advertising business. Concurrently, a recent Wall Street Journal report indicated that OpenAI CEO Sam Altman declared a "code red" to prioritize enhancements to ChatGPT's quality, potentially postponing other product initiatives, including advertising ventures.
