Is gender confined to two options only? Diverse genders have been a hot topic for discussion. Turns out there are more genders than we know. The latest categorization of such groups under LGBTQ+ has made it easier to identify genders more easily.
Non-binary communities are those individuals that don’t identify themselves as masculine or feminine. Their gender preferences vary from bigender, trigender or being gender fluid. They don’t seem to stick onto one category of sex preference but many.
A facial recognition to identify the non-binary would be a great step to a liberal society. Group of scientists affiliated with Harvard decided to create a facial recognition to identify gender minority subgroups such as LGBTQ+ and Non-binary communities that they claim to be ‘racially-balanced’ database that corporates the subset of LGBTQ+ people and ‘inclusive-gender’ database that mitigates classified gender biased algorithms. But however, there is a loophole in this system that proves fatal if not corrected properly says Os Keyes, an AI researcher at University of Washington.
According to Keyes, “The researchers go back and forth between treating gender as physiologically and visually modelled in a fixed way and being more flexible and contextual”. The fact that the model was not initially discussed with trans-individuals is in doubt.
Facial Recognition systems have always been considered problematic in its analytics of recognising faces, stating it tends to show bias towards race and gender, especially those who don’t identify under a single gender, it has always undergone repairs and changes in its workings. Several stores have taken down facial recognition as a form of identification. After the recent BLM has created a wave in the racial bias systems, consideration about processes and data used to in certain software are needed to be taken into consideration.
Despite of the risks underlying this feature, the creators of the gender unbiased facial recognition aims to improve their system performance by eliminating societal gender biased algorithms and lack of LGBTQ+ and non-binary communities identification under popular benchmarks. Without bringing out differentiation in the gender classification, there would be a false sense of progression in our society that needs to be eliminated.
Keyes, however, opposes the study saying it contains fixed and modelling the non-binary people is ‘third gender’ that lies between men and women which is not even the classification they fall under. Also non-binary people might change their identification and preferences in different forms and identifying them in all those forms that cannot be scientifically possible because little to no change could happen in their appearances. It also gives less security to binary and non-binary individuals who refuse to disclose their identity.
In a growing world, diversities form in every aspects of our life which is beautiful, but categorizing everything under analytical heads –is it a boon or bane?