Algorithms are especially valuable for the evocation of science based neutrality, because it's really just math. Silly liberals saying 2+2 isn't 4. This conversation is so Orwellian.
Numbers don't actually exist in the world. Quantifying is a thing *we do to something* - and we choose how to do it.
Numbers don't actually exist in the world. Quantifying is a thing *we do to something* - and we choose how to do it.
Comments
Then you feed that dataset into a robot. Guess what, you made a racist robot.
This is called algo-washing and it's pretty much as old as computers.
You can't just turn the tap and have Technology pour into your glass. There needs to be some kind of interface - a way to give inputs and get outputs.
This is not solely a commercial pressure, but the nature of the game - the ways scientists package their own tools is often extremely limited.
https://www.ebit.ks.gov/resources/governance/it-executive-council/itab
But different activities also attain different outcomes.
By making certain activities preferable to others, any tool (regardless of whether it's cutting edge IT or a new kind of pasta shape maker) changes the set of *possible outcomes*.
A gun kills many men before it's done
Hundreds
Long before you shoot the gun
Men in the mines to dig the iron
Men in the mills to forge the steel
Men at machines to turn the barrel, mold the trigger, shape the wheel
We can go deeper
But "algorithm" is really just maths. Algorithmics doesn't concern itself with form factors and applications.
In formal definition an algorithm is essentially any math software does. But it is also used to refer to the sort of black-box programs behind twitter or Facebook or chatGPT, which have all kind of hidden embedded biases.
But corporations and bigots know that, if they pretend it does, many of us will just let them slide instead of using precise language to call them out.
If you use "+" inside of software ICE uses to track deportation cases it's a whole deal entirely