Their findings, available on the preprint SSRN server, suggest that humans are more likely to believe that a robot has a "mind" if it has eye-like structures on its face than if it does not.