It's too early to worry about "paper click maximizers", but we're up to our neck in the problems of "profit maximizers", and all AI is good for reducing labor costs for the profit min-max-ers.
For actual AI problems, what's realistically going to play out in the next decade is more of the same. Computer doesn't detect PoC skin. Your eligibility for X score is 100 points lower because we detected your ZIP code belongs to a poor (aka majorly PoC) neighborhood. Another 100 points if you're a woman.
Why was I rejected? The clerk in call center says "Computer says no. I'm not allowed to override it, I'm just the clerk."
in a communist society we could probably work to counter racism in datasets as it is that would be expensive and capital values profit over anti-racism
It's too early to worry about "paper click maximizers", but we're up to our neck in the problems of "profit maximizers", and all AI is good for reducing labor costs for the profit min-max-ers.
For actual AI problems, what's realistically going to play out in the next decade is more of the same. Computer doesn't detect PoC skin. Your eligibility for X score is 100 points lower because we detected your ZIP code belongs to a poor (aka majorly PoC) neighborhood. Another 100 points if you're a woman.
Why was I rejected? The clerk in call center says "Computer says no. I'm not allowed to override it, I'm just the clerk."
Why is the data set racist? It was trained on a society that is fundamentally racist.
in a communist society we could probably work to counter racism in datasets as it is that would be expensive and capital values profit over anti-racism