Science Scavenger Hunt

Stanford researchers claim to have taught a machine to detect sexual orientation.

The controversial study, authored by Michal Kosinski and Yilun Wang, works by analyzing photos from dating profiles and detecting key facial features.

When analyzing five clear photos of a person’s face, the machine had an 83% accuracy for women and 91% for men in the study.

Though the intention of the study was to bring awareness to privacy risks and the danger of facial recognition software, according to the New York Times the authors of the study have received several death threats for their publication. Critics have called the study “racism by algorithm” and suggested that the theory behind it “is a mess”.

1 Comment so far

  1.   fuglsang on October 23rd, 2017

    This is an interesting piece, Nick, and I think you do a good job of focusing on the important pieces: the controversy.

    The one suggestion I would make is to repeat the “detect sexual orientation” angle in the second graf (your actual lead). Listeners may not have caught the entire throwaway lead, so don’t worry too much about repeating key points.

    At some point include attribution to your source because people (like me) may want to find out more.