Multiple examples are querying for "female" (which could be fine), and this prompted a thought...
What happens when a customer is searching for hiring purposes, and searches specifically for "male"? Or "young", or "unmarried", or "childless", or "straight", or "white", or "non-disabled", or "non-veteran"?
The data is out there, and is bought and sold heavily. (You mention that dog ownership is something you query over.)
What queries are you going to permit, and what not?
Even with current tools, it seems a lot of people people do this casually. Besides the many biases that people will openly admit on HN, I'm reminded of when someone told me to use one of the popular hiring sites to filter out candidates who weren't in early/mid-20s. (They spoke of it as if it was clever to use graduation year, since the site didn't let you filter by age directly.) Aaaannnndddd... the hiring sites surely have that search history information that recruiters and hiring managers for numerous employers are doing, unless they're intentionally discarding it against all data-appetite industry convention, so should be easy fodder for some energetic regulators/lawyers.
What happens when a customer is searching for hiring purposes, and searches specifically for "male"? Or "young", or "unmarried", or "childless", or "straight", or "white", or "non-disabled", or "non-veteran"?
The data is out there, and is bought and sold heavily. (You mention that dog ownership is something you query over.)
What queries are you going to permit, and what not?
Even with current tools, it seems a lot of people people do this casually. Besides the many biases that people will openly admit on HN, I'm reminded of when someone told me to use one of the popular hiring sites to filter out candidates who weren't in early/mid-20s. (They spoke of it as if it was clever to use graduation year, since the site didn't let you filter by age directly.) Aaaannnndddd... the hiring sites surely have that search history information that recruiters and hiring managers for numerous employers are doing, unless they're intentionally discarding it against all data-appetite industry convention, so should be easy fodder for some energetic regulators/lawyers.