OpenAI, the company behind ChatGPT, is exploring a new advertising model within its popular chatbot. The company has started testing visual ads in conjunction with its image generation feature, marking a significant expansion in its advertising approach. This development comes as OpenAI seeks to diversify its revenue streams, leveraging its massive user base of 1.2 billion weekly users.
The ads in question will not be integrated directly into the generated images but will appear separately with clear labeling as sponsored content. These ads can focus on products being used or services and experiences provided, giving users promotional visual content without merging it directly into their generated images. This approach aims to maintain a clear distinction between AI-generated content and paid advertisements.
A crucial aspect of this testing involves how ads are selected for display. OpenAI uses automated systems to analyze the context of conversations before deciding to show a specific ad. This means the decision to display an ad is not solely based on a user's engagement with the image generation feature but also considers the context of their conversation. This method raises questions about balancing personalized ads with user privacy, especially in conversations that may include sensitive information.
To ensure the safety and appropriateness of these ads, OpenAI has engaged with external companies, DoubleVerify and Integral Ad Science, which specialize in ad verification and brand safety. These tests are conducted in a controlled environment that does not rely on real user conversations, aiming to assess how well the ad system adheres to OpenAI's safety guidelines. This external review is critical as ChatGPT expands its advertising, with one of the main challenges being to prevent ads from appearing in inappropriate contexts.
The motivation behind OpenAI's move into advertising is largely economic. With ChatGPT serving a vast user base, the platform sees a significant opportunity to build its advertising business alongside revenue from paid subscriptions. While ChatGPT already displays text-based ads, the introduction of visual ads linked to image generation represents a further expansion of this model. This strategy allows the company to generate revenue from a broad user base, including those not subscribed to paid plans.
The introduction of these ads in ChatGPT has started in 31 European countries since late August, targeting users on free and Go plans, while users on Plus, Pro, and Enterprise plans do not see ads. This rollout occurs amid increased European scrutiny of ChatGPT's operations and its widespread adoption among users. The European Commission has classified ChatGPT as a very large online search engine, given its significant user base of 159.1 million monthly users in the EU.
The classification by the European Commission not only reflects the number of users but also imposes greater requirements on ChatGPT regarding transparency and disclosure of its advertising activities. By January, OpenAI is expected to publish information about its ads for the past year, including details such as the advertisers, the duration ads were shown, targeting mechanisms used, and the number of people reached. This level of transparency aims to provide users and regulators with a clearer understanding of how ChatGPT manages its advertising.
Key points
- OpenAI is testing ads in ChatGPT's image generation feature to diversify revenue streams.
- The ads will appear separately from generated images and be labeled as sponsored content.
- The testing is initially available in 31 European countries for users on free and Go plans.
