User:Sxologist/sandbox/Michal Kosinski

Michal W. Kosinski (born Michał Kosiński May 8, 1982) is a Polish-American psychologist and Associate Professor at Stanford University Graduate School of Business. Kosinski is best known for his research regarding AI, psychometrics, data privacy and facial recognition technology.

Career and research
Kosinski received his PhD in Psychology from the University of Cambridge (UK) in 2014. He was formerly the Deputy Director of the University of Cambridge Psychometrics Centre, a researcher at Microsoft Research, and a post-doctoral scholar at Stanford’s Computer Science Department.

In 2018, Kosinski published a letter to the Stanford University website saying that he had no involvement in the Facebook-Cambridge Analytica data scandal. He wrote that inaccurate press reports had confused him with Dr Alexandr Kogan, who worked at Cambridge University at the same time.

In 2019, Kosinski published a study which used neural AI networks to guess the sexual orientation of 14,000 white Americans taken from dating sites. In one test, the algorithm was shown two photos where one picture was definitely of a gay man and the other straight, and it guessed correctly 81% of the time. In another, this rose to 91% accuracy when the algorithm was given 5 photos. With women, it was correct 71% of the time. Kosinski reported this was the dangers of future technology which may put people's privacy at risk. He also cited this as evidence of the prenatal hormonal theory of sexual orientation; that hormonal fluctuations or receptivity to them in utero may be a dominant factor in determining a man or woman's sexual orientation, as exposure to testosterone in utero influences facial morphology. Kosinski was criticized by LGBT activist groups HRC and GLAAD who demanded that Stanford denounced it as "junk science". Kosinski defended himself and called the criticsms "knee jerk", and that "scientific findings can only be debunked by scientific data and replication, not by well-meaning lawyers and communication officers lacking scientific training".

According to Google Scholar, Kosinki's works have been cited more than 9,000 times and he has an h-index of 39.