Folks who have any language besides English and any writing experience in any field on your online CVs on indeed and places like that, we get bombarded with AI “editing” and “proof reading” and even “copywriting” posts – in Canada too
Everyone thinks these exploited workers are far away when the truth is Scale AI - dba https://Outlier.ai - is employing US citizens, failing the economic realities test, and misclassifying workers who should be W2 so they can pay a pittance while raking in profits in the form of wage theft.
Of course no one has heard about this. Because AI is cool and neat and it isn’t stealing and it isn’t exploitation and it will usher in a Golden Age. There are no downsides. Haven’t you heard that?
Yeah. There was coverage early on in the Atlantic, but from then on people seemed mostly focused on the theft & environmental impact rather than the horrendous labor exploitation issues. 🫠
? Kate Crawford has written about this extensively for years. Many articles out there on barely paid human intelligence labeling images for AI. Content moderation..
Not mainstream tho well documented and more recently CO2 / energy footprint.
I was assigned a mentor at work who is the head of the AI initiative. He is an impressive person, but I perceive AI to be the outsourcing of factual research. Combine that with a national news infrastructure that freely distributes disinformation. Not good.
comes up when talking to folks about EVs too. they will tout that they consume less CO2 from the consumer side, but forget the massive amount of fossil fuels and water they consume in their manufacture.
people love to think they’re getting something for nothing
I wonder if half of it isn't also leftover from the soda industry's marketing efforts to shame littering and pass the blame of recycling onto the consumer. Like it's a perspective extremely focused on "my own personal consumption is the most responsible for pollution"
For what it’s worth - it depends on the company. In the veterinary world, we outright pay normal veterinarians to assist with this process. And the AI helps do trivial, crappy tasks they don’t want to do anyway (or can easily make mistakes on).
Not all AI are LLMs and they are not all unethical.
Probably because it’s not ubiquitous enough anymore - the LLMs largely can self label their own data and humans in the loop just give it a thumbs up or down.
They say “Yes, the model correctly counted the number of blood cells on that slide” or “no it didn’t and here’s why” vs. counting every slide themselves and labeling it to be used as training data.
Or yes it recalled this fact without bias, etc. General validation.
LLMs are trained on unlabeled data. Comparing working conditions of white-collar workers, whose jobs only exist on a computer, to sweatshop workers is beyond ludicrous.
My mom mentioned this after seeing the recent 60 Minutes story, and I went on a nigh 10-minute infodump rant about the shady shit these companies pull.
Comments
Not mainstream tho well documented and more recently CO2 / energy footprint.
Definitely more visibility needed!
we need to start from the top and work our way down
comes up when talking to folks about EVs too. they will tout that they consume less CO2 from the consumer side, but forget the massive amount of fossil fuels and water they consume in their manufacture.
people love to think they’re getting something for nothing
it will make my students brain-dead
exploit these workers + steal writers' work
and burn up the planet
Not all AI are LLMs and they are not all unethical.
Or yes it recalled this fact without bias, etc. General validation.