For the first time, we’ve given external researchers a way to study AI’s impacts using real, privacy-preserved Claude usage data. To date, this work has only been possible within AI labs. We can’t tell the whole story alone, so we opened up our tools.
https://t.co/suLJcARAu1
Three research groups—Stanford’s Social and Language Technologies lab, Oxford’s Human Information Processing Lab, and METR—designed independent studies to analyze the aggregated outputs from 250,000 https://t.co/SVzAB0InAs or Claude Code conversations between April and May 2026.
The SALT Lab studied how people collaborate with AI. They found that over half of these conversations involved consequential tasks—work that affects other people or is hard to undo.
Read their full writeup here: https://t.co/bZ7Qu5nZjb
The other two studies are ongoing: HIP Lab is studying how Claude's behavior relates to how people feel when using AI, while METR is estimating real-world productivity gains from coding agents. We'll share more from both soon.
AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
guardrails — The checks around a model that block bad inputs and outputs — filters, validators, and permission rules the model itself can't override.
Why it matters
Independent measurement of real AI agentAn AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.Full definition → usage is leaving the labs. The first result — over half of sampled Claude and Claude Code conversations involved consequential, hard-to-undo work — sets the stakes for guardrailsThe checks around a model that block bad inputs and outputs — filters, validators, and permission rules the model itself can't override.Full definition →, and METR's productivity study is coming.
