The Rise of 'Synthetic Respondents': Can You Be Cloned for Research Purposes?

The 'synthetic respondents' industry—digital avatars for surveys and research—is raising hundreds of millions of dollars, but studies show they are more extreme and less diverse than real humans, and one company already got the US election forecast wrong.

The article examines the 'synthetic respondents' industry—digital avatars designed to replace humans in surveys and scientific experiments. Companies like Simile (from Stanford University) and Auru (from MIT) have raised hundreds of millions of dollars, promising to predict consumer and political behavior. However, Auru predicted a victory for Kamala Harris in the 2024 US election—a forecast that turned out to be wrong. Studies show that synthetic respondents tend to be more extreme and less diverse than real humans, and sometimes statistical correlations are reversed. Dr. Neil Zur, founder of Askit, criticizes the field but believes in an alternative approach based on cognitive patterns.