Gaydar research at MIT and the state of LGBTQ+ rights around the world with technology

Gaydar research at MIT and the state of LGBTQ+ rights around the world with technology

An important study related to "Gaydar" at MIT was conducted by Carter Jernigan and Behram F.T. Mistree in 2009. The study analyzed 4,080 profiles collected from MIT's Facebook network and proposed a method for predicting a person's sexual orientation based on the composition of a user's friendships. Specifically, the study showed that if a user's friends had more men who self-identified as gay, they were more likely to be gay themselves.

Using a logistic regression model, they showed that a user's sexual orientation can be inferred with considerable accuracy from friend list data alone, raising a serious privacy issue: "Online personal information can reveal attributes that a person has not disclosed. In other words, there is a risk that sensitive attributes can be read off from the structure of the surrounding information alone, even if the individual does not explicitly state them themselves.

Also in 2017, research by Michal Kosinski and Yilun Wang of Stanford University made headlines. They experimented with using deep learning to predict sexual orientation from facial photos and reported that they were able to determine whether a person was gay or straight with 81% accuracy for men and 74% for women. While the study attracted technical attention, it also provoked strong criticism from an ethical standpoint.

The MIT Media Lab is also working on the "AI Comes Out of the Closet" project, which aims to use AI to support LGBTQIA+ advocacy efforts, using AI interaction and virtual characters to help people practice the social skills needed for advocacy in a safe environment. to practice the social skills needed for advocacy in a safe environment.

These technological developments are increasingly complicating issues of personal privacy and ethics. In particular, given that sexual orientation is a subject of significant persecution in some countries, careful application of technology is required.

In fact, as the UK's "The Guardian" reported in 2013, homosexuality was illegal in about 40% of the world's countries at the time. In particular, homosexuality was criminalized in 41 of 53 Commonwealth countries, and many of these laws are remnants of colonial-era British law. Even if homosexuality is not legally prohibited, social prejudice and violence are serious problems in some countries.

For example, in Jamaica, a tragic incident was reported in which 17-year-old Dwayne Jones was attacked and killed by a mob for "cross dressing. In this context, "cross dressing" refers to the act of dressing in a gender that is different from the socially prescribed gender. For example, it refers to men dressing as women or women dressing as men and is a form of self-expression. In some societies, however, these expressions are subject to intense prejudice and violence.

Thus, in an age when technology has made it possible to infer sensitive personal information, there are serious risks involved that go beyond mere technological curiosity. In countries or regions where sexual orientation is illegal or poses significant social risks, the assumption of attributes against one's will can create life-threatening risks.

MIT's "Gaydar" study vividly highlights how complex the intersection of technology, privacy, ethics, and the global human rights situation is. Addressing these challenges will require a comprehensive approach, grounded in ethics and human rights, that brings together technologists, policy makers, and civil society.