| I've built one of these before and the devil is so very much in the details, but really even more in the context. There's the easy stuff: "ninja" for example. Nobody, in the year 2023, should be using that word in a job post. But then there's the hard stuff that really matters. If you want someone who has experience being a "white hat" hacker, that's perfectly alright, but if you want "white people only" well that's obviously bad. If you don't flag the latter you look like a joke, but if you flag every instance of the word "white" then it feels overbearing and like the tool isn't very smart. I'm pretty sure these will never actually be useful on a superficial word-matching basis. They need to look at broader phrases and context. And then there's the real problem that even if the tool helps the hiring manager / recruiter sweep their inbuilt biases under the rug to get better applicants, they're still the ones making the hiring decisions. |