DEV Community


Discussion on: Fool proof women only verification for a platform, how would you solve this?

dmfay profile image
Dian Fay

As far as I understand it facial recognition technology is roughly at the party-trick level, able to guess not quite as well as humans can, and we ourselves are hardly infallible. Algorithms so far have also tended to display training bias and are worse at categorizing women and people of color, to say nothing of trans and gender-nonconforming people. It's completely unacceptable for something this important, but to my point elsewhere no automated general-purpose solution is acceptable for what fundamentally isn't a technical problem.