Based on a single image, the algorithm was reportedly successful at distinguishing a man’s sexuality 81% of the time and a woman’s 74%. The stats increased to 91% and 83%, respectively, when provided five pictures.
Naturally, the study received widespread and immediate criticism. Academics and people within the LGBT community alike took immediate issue with the study’s claims.
The study, penned by, Michal Kosinski and Yilun Wang, both puts LGBT people in danger and also reduces sexuality to appearance. They also claimed their findings show ‘strong support’ for the idea sexuality stems from hormone exposure in the womb.
Now, however, the Kosinski and Wang’s work is under ethical review.
The Outline reports the Journal of Personality and Social Psychology is reviewing the study following the intense backlash.
An editor for the journal, Shinobu Kitayama, said the review is happening now and results will be announced in ‘some weeks’.
Unsurprisingly, Kosinski and Wang are on the defensive. Late last week, Kosinski tweeted people’s opinions would be ‘stronger’ if they ‘read the paper and our notes’.
Given the implications of the study and subsequent responses to it, a review was only a matter of time. The review’s findings could be serious for the future of AI technology, academia, and the LGBT community.
No comments:
Post a Comment