New research reveals facial recognition softwear misgenders transgender and nb people

I don't see how an AI could ever do more than classify by sex, with a second layer for gender nonconformity - at the more obvious ends guess at trans, but in between there simply isn't any physical distinction between 'agender, genderqueer or non-binary' - or people who don't give a stuff about gender and don't conform to stereotypes.

It's a complete and utter nonsense.

More troublingly, those who identified as agender, genderqueer or non-binary, were misclassified 100 percent of the time because these gender identities have not been built into the algorithms.

This is terrifying. What kind of person thinks that facial recognition software would or should ever be used to assume identity? What next - an algorithm that uses the length of your nose to predict political affilation?

I'm not sure I'm a fan of facial recognition tech in general, I can see it has its uses but a lot of downsides so not too bothered if wokeness kills it.

Reality is a T*RF.

We have had reality for so very many millennia that it is actually hardwired into our human brains now to instinctively be able to tell men and women apart. Amazing, isn’t that. Tiny little children can do it.
Are they T*RFs too?

And what about us that id at times as a cat or a table? How will the software deal with that, other than maybe breaking down from confusion

If tech is terf then all the moaners need to step back from the tech surely and go and create their own stuff.

Will be interesting to see if any of the main media outlets give this any mention. That would wake up loads of people.

The programme correctly identifies a person's sex. Like most of us do with our own eyes.

I don't see the issue. Like pp said, how can you make it identify a feeling that is purely subjective and changeable. It would make it pointless.

Well, the software is clearly exceptionally clever!
Let’s face it, it’s doing what it is supposed to. It is identifying people correctly.

Something tells me on travelling days he always knows who he will be.

I nearly wrote something like that in my post.

Facial recognition software which doesn't ignore external gender signifiers would, I think, be problematic for people during real transition.

More troublingly, those who identified as <a class="break-all" href="http://go.mumsnet.com/?xs=1&id=470X1554755&url=www.psychologytoday.com/us/blog/sex-sexuality-and-romance/201807/guide-genderqueer-non-binary-and-genderfluid-identity" target="_blank">agender, genderqueer or non-binaryy^y, were misclassified 100 percent of the time because these gender identities have not been built into the algorithms.

How would one do that? Look for the duper's delight sneery smile so many of them do?

The piece doesn't say what the hit rate for transwomen was - I'd guess fairly high because an AI may well be able to detect sex Male

I would say males who identify as trans were deliberately left out of the piece, because biology is a terf and the high identification rates are probably op hurtful.

Tangentially related thread:

www.mumsnet.com/Talk/womens_rights/3732308-Future-of-Legal-Gender-academic-talk

Why are they being allowed to influence policy for all?

Some of these people are legitimately batshit crazy.

Errol Excellent point re the passport photo.

Something tells me on travelling days he always knows who he will be. (I've chosen his preferred pronouns before anybody points the Donald Sutherland finger).

This is a satire, surely?! Or maybe not, but shows how crazy this is. Of course gender cannot be recognised because it is a FEELING! Do the algorithms show favourite colour, preference for Bach or Mozart, etc? Gender identity is no different. It is internal in some, entirely absent in others. How on earth could it be detected?

More troublingly, those who identified as agender, genderqueer or non-binaryy*, were misclassified 100 percent of the time because these gender identities have not been built into the algorithms.

The training data for the AIs presumably doesn't have these nebulous categories marked. And if they did - and they were present in statistically significant numbers - then we'd probably just find lots of women and some men being miscategorised.

The piece doesn't say what the hit rate for transwomen was - I'd guess fairly high because an AI may well be able to detect sex Male + external 'feminine' markers such as makeup, and categorise them as transwomen.
I'd be curious to know if the hits and misses for transmen were broadly beard or no beard...

Yes Ereshkigal everything is so problematic these days!

Technology is a terf.

Very relieved to hear that the technology does not yet exist to read people’s minds

Surely humans can't be that much more stupid than the average warthog?

Not all of them anyway

You and me both, Jane. People didn't "guess people's gender". They, as people still do now, perceived their sex. No one gave a flying fuck about "gender" in the sense the writer means.

That picture in the Wings over Scotland link of the group with their axes and hammers and baseball bats. What a shit shower.

Quaint idea from the 90s?

I'm no that old but I do remember the 80s and we could correctly sex people by looking at them back then too, despite the fashions of the time being far less sexist stereotypical than today.

I'm willing to bet good money that humans have been correctly sexing one another throughout our history. I mean, other mammals manage to figure out which of them is male and which female and they don't even wear clothes and make up. Surely humans can't be that much more stupid than the average warthog?

Anyway, this piece is massively transphobic, missing the point that gender isn't about looks, it's about how someone feels isn't it?

"these systems have an extremely limited view of what gender looks like.”