It’s not really useful to talk about #skynet and robot overlords when discussing the dangers of #artificialintelligence because that’s not what we have to worry about, at least not for a while yet.
It is much more useful to talk about the concept of #policymurder which is when systems codified by law and regulation squeeze people into destitution and death. See also “social murder.”
@carnage4life Bishop William J. Barber explained policy murder in a June 2022 HuffPost op-ed: we have made certain choices about our laws and administrative procedure in the United States that create widespread poverty and that poverty kills roughly 700 people every day.
@carnage4life Less sophisticated algorithms have already been making decisions like these for years in some places, using biased (by which I mean usually horribly racist) training data from past decisions made by humans.
The #policymurder by algorithm has been going on for years and the policy murder by bureaucracies and institutions has been going on for decades or centuries before that.
You can automate most or all the grinding bureaucratic institutional processes to support policy murder with artificial intelligence.
And you can have less accountability. Go from “I was just following orders” to “we were following procedure” to “the AI did it so none of us are at fault.”
The danger of #artificialintelligence is not that it’s going to launch a global nuclear war it’s that it will decide to deny you health insurance or to send armed police to your house or that you shouldn’t get benefits you desperately need because some humans decided it should work that way.
And/or because it’s sloppily implemented by the lowest bidder but no one cares because the ones who get hurt are poor people.
Add comment