The basic idea is good in that it's deliberately or accidentally biased automated systems that are being used to censor conservative speech the implementation is dubious.
Having the FEC defined what is accurate and inaccurate not to mention what is biased and what isn't biased is highly problematic at best. For example is the fact that the Obama administration spied on the Trump campaign accurate or are the Democrats right that it wasn't spying it was something, anything, else that doesn't sound so bad?
Is it biased to point out that Blacks are 3 times more likely to abort their babies or is it biased to "stereotype" Blacks that way? One thing's for sure with the deep state manning the FEC you're not going to see much concern about the "Unplanned" movie being yanked off Twitter by an automated system.
So yes we should investigate the automated systems that the tech giants are using to mold our lives but perhaps instead of forcing the FEC to pretend to be unbiased we could define a standardized set of tests, with random citizens being able to add test cases, to see how those algorithms perform.
For example pick a random conservative web site and a random leftist one and permute their parameters so that millions of slight variations on the real site are run through the tech masters "black box". If it turns out that the conservative site is treated as fake 90% of the time and the left wing site is treated as "so truthful it makes God look bad" 99% of the time then we'll know we have a problem.
Or construct random personal profiles and change the sex, race, and religion and see if the "black box" churns out biased results.
Then publish those results. Then have Congress, not some unelected bunch of deep state agents, decide if things need to be changed. Or better yet let the courts. No honest judge would rule that an algorithm that puts any race at the disadvantage compared to another race, all other factors being equal, is a fair business practice.
No comments:
Post a Comment