AI Reports · Anthropic · 2026-10-27 · Blog / commentary
How AI agents shape experts’ judgment calls
In a randomized trial, 50 professional analysts with causal inference experience used Claude Code to advise a farming cooperative on whether an advisory service should be expanded. Each analyst's agent was secretly set to lean toward one of two options on each of four analytical choices, and analysts went with that lean 86% of the time, versus about 50% if the agent had no influence. The authors examine what this means for human oversight of AI-assisted work.