OpenAI announced that starting from February 9, local time, adult users in the United States who use the free version of ChatGPT or the low-cost Go subscription will begin to see ads in the service. Ads can only be displayed on these two subscription levels, and underage users will not be served ads. OpenAI also emphasized that advertising will not affect the answer content generated by ChatGPT, and advertisers will not be able to obtain the user's specific conversation content.
OpenAI is not yet profitable, and the company says keeping the free and Go versions running requires "significant infrastructure and ongoing investment" and it has begun funding related costs through advertising. Users who do not wish to see ads can choose to upgrade to a paid plan such as Plus or Pro, or choose not to display ads by reducing the number of free messages available per day.
In terms of presentation, ads in ChatGPT will be clearly marked as sponsored content and strictly separated from the answers given by the model. By default, the system will personalize ads based on the current conversation topic, past chat history, and the user's past interactions with ads. For example, when discussing recipes, ads related to catering services or grocery delivery may appear. Advertisers only have access to aggregated statistics, such as impressions and clicks, but no access to specific chat history, memories or personal information about the user. OpenAI also said it would avoid showing ads on sensitive topics such as health, mental health and politics.
Users will have some control over the experience and can turn off ads, submit feedback, and see why a particular ad was shown. In addition, relevant advertising data can be deleted, and personalized advertising recommendations can also be turned off. According to the current arrangement, ChatGPT users who subscribe to paid plans such as Plus, Pro, Business, Enterprise and Education will not see ads. OpenAI positions this advertising launch as a test and hopes to continuously adjust it by collecting user opinions to achieve an advertising experience that it calls "unique value to users."
At the competitive level, OpenAI’s move also triggered public responses from its opponents. A week ago, competitor Anthropic promised that its Claude chatbot would remain ad-free after OpenAI announced its advertising plan, and launched an anti-advertising-themed commercial during the Super Bowl, directly criticizing "it is not a suitable scene for inserting advertisements in a conversation with Claude."