RE: https://mstdn.social/@amydiehl/116343709194773175

AI is Bias at Scale.

AI is making a product of "We've always done it this way."

@tinker I actually tried an… “experiment” last year.

I sent two copies of my resume in for a bunch of jobs. Lots of different jobs, but a new account for each resume for all ten. Marked myself as male for the demographics for the male resume, “prefer not to answer” for the femme version.

One had my name and my dates and so on so forth. Full first name to signal “male”.

Other had shortened first name (is genderless, if more commonly male). No dates. Same “things” there, but just less info that made me seem “male”. Less detail about some of the work I’ve done, more high-level. Literally, I applied my (admittedly dated) linguistics skills to it, to the point that I asked my dad to review “a resume” for me and he said “wow this chick sounds a lot like you”.

Male-coded resume got a callback from every job.

Femme-coded? One call out of ten applications.

And it was not a bad resume, it read to my dad as “pretty much exactly you, wow, you guys should hire her”.

I only interviewed with the one who responded to the femme-coded resume. I did not, unfortunately, get the job, though I _did_ get a short “oh I thought you were a chick” look from the hiring manager when we zoomed, which I will forever be proud of.

Edited to wrap “experiment” in scare quotes to make it more clear that I am not a professional jobologist surveying the mating habits of the great North American job-beast.

@b4ux1t3 @tinker Why did you remove dates?

@msbellows @tinker just to give no indication of age at all. You could piece together that the person is in their 30s, probably, but by my estimation the range is from like 29 to 41.

This was during a time that I wasn’t _actively_ looking for a job, just kind of testing the waters. Goal wasn’t to get hired but to see how bad the market is for me, in specific, and “hey why don’t I run a run experiment”.

It was more depressing than fun.

@b4ux1t3 @tinker Yeah, I'll bet. Hoo boy.