What a doozy.
New chatGPT revelations.
Kenyan workers paid less than $2/p/h were required to read through large tracts of harmful and abusive content, in order to label and teach the AI what was bad:
“To get those labels, OpenAI sent tens of thousands of snippets of text to an outsourcing firm in Kenya, beginning in November 2021. Much of that text appeared to have been pulled from the darkest recesses of the internet. Some of it described situations in graphic detail like child sexual abuse, bestiality, murder, suicide, torture, self harm, and incest.”
The hiring company to which this work was outsourced claims to be about ‘ethical AI’ and boasts of having lifted people out of poverty. But what were the actual ethics and harms considered in doling out this sort of lowly paid, trauma inducing work?
I’m returned once again to the sombre thought that the digital world and its markets, for all its hype about automation, AI, and delivery at the press of a button, is absolutely still reliant on the exploitation and labour of ordinary flesh and blood. It’s the case whether we’re considering content moderation or your poor local courier driver trying to put food in the table with umpteen contracts.
It’s the dank Victorian factory, still. But it’s everywhere, all at once, now.
