MIT Technology Review reports that AI companies such as Anthropic and OpenAI regularly release reports describing how people use products like Claude and ChatGPT. However, AI researchers point out that these companies only publish the data they choose to make public. According to Anka Reuel, a computer science PhD candidate at the Stanford Trustworthy AI Research group, there is no independent source available to corroborate the figures the companies release.
Why it matters
Understanding how AI tools are actually used shapes public debate, research, and policy. If the only available information comes from the companies themselves and cannot be independently verified, the broader picture of real-world AI use remains incomplete.
Who should care
Researchers, policymakers, and journalists who rely on usage data to assess the impact of AI systems have reason to weigh the limitations of company-published reports.