William Sims Bainbridge — co-director of Human-Centered Computing at the National Science Foundation (NSF) — wrote in a review paper published in last week’s Science that virtual worlds would be ideal for performing sociological experiments.

Cited benefits include tapping into an unusually large population of subjects, acquiring lots of quantitative data about everyday activities, and repeating experiments under near identical circumstances.

A few pioneering researchers, such as Mary Lou Maher at The University of Sydney in Australia, are also using virtual worlds as a test platform for artificial intelligence, observing how a given AI program interacts with real people.