OpenAI has initiated testing of advertisements within its ChatGPT platform for users in the U.S. who are on the Free and Go subscription tiers. The Go plan, introduced globally in mid-January, is a lower-cost option priced at $8 per month in the U.S. In contrast, subscribers to OpenAI’s premium plans, including Plus, Pro, Business, Enterprise, and Education tiers, will not encounter any ads, as confirmed by the company.

Addressing potential concerns regarding the impact of ads on user experience, OpenAI released a blog post stating that advertisements will not influence the responses provided by ChatGPT, and user conversations will remain private from advertisers. The company's objective behind this move is to facilitate broader access to more advanced ChatGPT features, while simultaneously upholding the trust users place in the AI for sensitive and important tasks.

This decision, first announced by OpenAI last month, faced strong criticism from rival AI company Anthropic. Anthropic launched a series of Super Bowl ads that humorously depicted how poorly integrated ads could disrupt the consumer experience in AI chatbots. These commercials featured glassy-eyed AI chatbot actors delivering advice interspersed with poorly targeted advertisements, directly mocking OpenAI's impending ad integration.

OpenAI CEO Sam Altman reacted strongly to Anthropic's jabs, publicly labeling their advertisements as "dishonest" and describing Anthropic as an "authoritarian company." The concept of advertisements within AI responses has generally met with user resistance; OpenAI previously encountered a backlash late last year during tests of app suggestions that users perceived as unwelcome ads.

Despite consumer apprehension, OpenAI is driven by the necessity to generate revenue from its widely popular chatbot. This revenue is crucial for offsetting the substantial costs associated with developing its advanced AI technology and expanding its business operations. While the need for revenue is understandable, critics remain concerned that advertisements could potentially bias ChatGPT’s answers.

OpenAI has countered these criticisms by asserting that ads will be optimized based on what is "most helpful" to the user. The company also guarantees that all advertisements will be clearly labeled as sponsored content and distinctively separated from the organic conversational output. In its testing phases, OpenAI has explored various methods for matching ads to users, considering factors such as the subject of their current conversations, their past chat history, and previous interactions with advertisements. For example, a user researching recipes might be shown ads for grocery delivery services or meal kits.

Furthermore, OpenAI has outlined stringent measures for user data privacy and control. Advertisers will not be granted access to individual user data, instead receiving only aggregate information pertaining to ad performance, such as total views and clicks. Users will also be empowered to review their history of ad interactions and clear it at their discretion. Additionally, users can dismiss specific ads, provide feedback, understand the rationale behind why a particular ad was displayed to them, and manage their ad personalization settings. The company has also committed to not displaying ads to users under the age of 18, nor will ads be placed in proximity to sensitive or regulated topics, including health, politics, or mental health-related discussions.