“Anthropic and OpenAI release curated reports on how users interact with Claude and ChatGPT, but independent researchers have no way to verify or challenge these findings. Stanford PhD candidate Anka Reuel highlights that no independent data source exists to corroborate company claims. This lack of transparency raises serious questions about whether self-reported usage data can meaningfully inform AI policy or safety research.”
Key Takeaways
- Anthropic and OpenAI self-select which user behaviour data they publish publicly, with no external audit process.
- Stanford researcher Anka Reuel confirms there is no independent data source to verify AI companies' usage claims.
- The absence of third-party corroboration limits the ability of policymakers and safety researchers to act on reliable evidence.
AI giants publish usage reports, but researchers say the data is hand-picked and unverifiable.
trending_upWhy It Matters
When AI companies control the narrative around how their products are used, it creates a significant blind spot for regulators, academics, and the public trying to assess real-world harms or benefits. Safety researchers cannot build accurate risk models without honest usage baselines, meaning policy decisions may rest on incomplete or flattering portrayals of AI behaviour. This dynamic also gives well-resourced labs a structural advantage in shaping public perception and regulatory outcomes. Watchdog bodies and governments exploring AI oversight frameworks should treat this data gap as a priority — mandating independent usage audits could be a critical next step.
FAQ
Why don't AI companies just release all their usage data publicly?
Companies cite user privacy, competitive sensitivity, and security concerns as reasons to limit disclosures. However, critics argue these justifications also conveniently allow firms to highlight favourable findings while omitting problematic patterns.
Who is calling for more transparency in AI usage reporting?
Academic researchers like Anka Reuel at Stanford's Trustworthy AI Research group are among those raising the alarm. The broader AI safety and policy research community has increasingly flagged self-reported data as insufficient for rigorous independent analysis.
What would independent AI usage auditing look like in practice?
It could involve third-party access to anonymised interaction logs under strict privacy controls, similar to how financial auditors access company records. Regulatory frameworks in the EU under the AI Act may eventually require such mechanisms, though enforcement details remain under development.
