It does not vary wildly between countries. The overwhelming majority of countries fall into the range of 10-30% of tech roles filled by women and the remaining 70-90% by men. I don't think that a wide variety of countries falling into a band of ~20% is what most people would characterize as "varies wildly".

Regardless, the cause of the disparity is not relevant to the discussion at hand. The real world population of people involved in technology is disproportionately of one gender, so an unbiased machine learning model meant to associate terms with genders will associate technology terms with men at higher rates. Sure, some may claim that the reason why disproportionately more men go into technology is because society is biased. But the AI's role was not to model a hypothetical world with no biases. Its role was to model the real world.

As I wrote in my root comment, some may see value in biasing their models to achieve results that fit what they see as an ideal worldview. But the point remains: to do this is to deliberately introduce biases, not remove them.

You don't think that the difference between a 1:9 ratio and a 1:3 ratio constitutes "wide variety"?

That's a proportional measurement. A small difference in an already small value can result in large proportional differences even when the absolute difference is small. This is sort of like how small towns might see their murder rate skyrocket by 300%, but in reality what happened was that there was one murder last year and three murders the following year.

Sure, but 30% of the population is not a small value, either.

Nearly all countries have rates of women in tech in a band between 10 and 30%. No country has a majority female tech workforce. The notion that there is a "wild variation" in the rates of women's representation in tech is simply not true.

I’m not sure how this happened, but somewhere along the line the claim that there are no inherent differences between genders became less extreme than the claim that there might be some. You’re not making the former claim, but I see it all the time and it’s just as nuts as saying that 100% of observed difference are due to biology.

The difference in treatment is probably because these days the people who talk loudest about the "some" are often trying to use it as a cover to support gender-essentialist defenses of an unequal status quo, while people who talk about "none" are obviously inaccurate in some ways but are for the most part harmless.

Well, I don’t know about you but I like my claims to reflect reality and not some ideal that is touted as “probably wrong but mostly harmless!”