Microsoft introduced Real-Time Crowd, an Application Programming Interface (API) that can accurately identify human emotions and reactions based on facial expressions during the 2016 Republic and Democratic National Convention
According to Microsoft's blog, Real-Time Crowd can identify eight emotional states including fear, contempt, anger, happiness, sadness, and surprise. The software features tools that connect to Microsoft's cloud computing services. Using various web applications, the data collected by the API will be analyzed by Microsoft's servers.
Microsoft's sales representatives noted that the technology tracks the emotional impact of entire groups of people simultaneously. The data can be used to study crowd response. It could identify stern faces of dissidents or angry faces of protestors in just a few seconds. Microsoft noted that the technology can be used to help blind people read facial expressions as well.
Microsoft's Real-Time Crowd platform was demonstrated at a DNC exhibit. It featured a small camera that took images of human faces, assigning the emotional profiles to every person in the crowd.
Meanwhile, digital rights activists and civil liberty groups warned that the technology is posing a threat to privacy.
In a 2013 press release, Electronic Frontier Foundation Staff Attorney Jennifer Lynch wrote: "Face-recognition technology is among the most alarming new developments because Americans cannot easily take precautions against the covert, remote, and mass capture of their images."
Facial recognition technology has long been used in secret by the government to identify criminals and individuals with questionable legality. Last month, a report by the Government Accountability Office (GAO) revealed that the Federal Bureau of Investigation has access to over 411.9 million images on its face-recognition database. The majority of these images are from people with no criminal record.
To diffuse privacy issues, the API requires developers to get the consent of the people whose facial profiles are being processed.