Face Values

How much can we really know about people from their faces?

A presentation about the perils of A.I. and facial recognition could easily fall into fearmongering clichés about the end of privacy and the dangers of evil corporations. Instead Face Values: Exploring Artificial Intelligence, an award-winning exhibit showing at the Cooper Hewitt in New York through May 25, humbly poses the question: How much can we really know about people from their faces?

A brief history of attempts to glean information from facial topography by criminologists, beauticians, doctors, and eugenicists primes visitors with a sense of the fallibility of the science. Then interactive stations allow them to watch tools designed to do things like detect emotions or assess age succeed or fail in real time.

Face Values offers moments of humor—with both state and corporate hubris as the butt of the joke—leavened with occasional horror at the potential human cost of releasing unripe systems built of shaky or biased data into the wild. Repeated reassurances from the curators that the personal information collected from museumgoers won't be stored or shared add a delicious ironic twist.