What 81,000 people want from AI
Summary
Anthropic's study aggregates open-ended responses from 80,508 Claude users across 159 countries, outlining what people want AI to do (boost productivity, manage life, learn, and create) and what concerns them (unreliability, job displacement, autonomy). It frames AI as a spectrum of benefits and harms—'light and shade'—that varies by region and occupation. The findings have implications for AI design, ethics, and deployment in SMBs seeking to automate processes while protecting workers and data.