The AI Observatory, a new research platform co-led by Stanford PhD candidate Anka Reuel, has launched to provide independent data on how people use generative AI models like Claude and Gemini. The platform aggregates real AI conversations collected with user consent from seven datasets, aiming to fill gaps left by major AI companies’ selective reporting, according to technologyreview.com.
AI companies such as Anthropic and OpenAI publish usage reports focusing mainly on work-related applications, but the AI Observatory found that AI use varies widely across models and includes many sensitive behaviors not reflected in corporate data. Reuel said the Observatory’s public platform offers researchers and policymakers a more comprehensive view by analyzing conversations beyond productivity, which are often filtered out in existing indices like the Anthropic Economic Index.
This independent data source matters as stakeholders make critical decisions about AI’s risks and benefits based on limited information. The Observatory’s findings highlight that personal and non-work uses of AI are more prevalent than previously reported. By revealing these broader usage patterns, the platform challenges the narrower focus of widely cited AI usage metrics and could influence future AI policy and research.
The AI Observatory’s aggregated data sets and analysis tools are publicly accessible, providing a resource for ongoing study of AI interaction trends. The platform’s launch marks a step toward transparency in AI usage, addressing concerns about the lack of independent verification of AI companies’ usage claims, technologyreview.com reported.