18 August 2026

Researchers launch platform to track what AI companies actually hide

• Anthropic, OpenAI, and other AI firms release only curated data about how people use their systems, obscuring real patterns.
• AI Observatory, a new public platform, analyzes unfiltered conversations to show what companies' reports leave out, including health advice and harassment.
• Nearly half of Claude conversations were filtered out by Anthropic's methodology, suggesting major gaps between published usage reports and reality.

How it was covered

The AlgorithmMIT Technology Review

Researchers launched the AI Observatory, a public platform analyzing real AI conversations to fill the gap left by companies like Anthropic and OpenAI, which only release curated data. The project found that major companies' reports significantly undercount sensitive uses like health advice, harassment, and sexual content, with nearly half of Claude conversations filtered out by Anthropic's work-focused methodology.