Millions Of Wasted Lives: A Hidden Army of AI Workers Are Fueling The AI Boom
AI Tasker: “I really am wasting my life here if I made somebody a billionaire and I’m earning a couple of bucks a week …”
In the space of just a few months, I’ve seen heinous, appalling AI art evolve into mediocre AI imagery with some potential uses. Dall-E just accidentally captured the perfect profile and expression for this article header image. However, there’s still a lot wrong with it. I lack the patience, knowledge, and ability to “make it better.” I also lack the desire to pay for more “credits.”
It’s “good enough.”
A few months ago, I got nowhere with suggesting to AI art developers that they involve actual artists in the designing of the programs.
I said, “There are a few things we learn in art school that the programs don’t seem to understand.” By “a few,” I meant hundreds — likely thousands — of concepts and techniques that we humans learn in art school (or some of us that are called ‘talented’ know instinctively).
As many have said, “AI art will be seen as ‘good enough’” thereby — so they say — putting real human artists out of business.
So far, here is what all AI is — visual, computational, verbal, technical, medical: a gigantic “Magic 8 Ball” with billions of facets.
This pixel here. Or, one to the right. One to the left.
Ages ago, I asked ChatGPT to write a poem about the sun setting over the Gulf of Mexico. It rhymed. It was as unoriginal and stereotypical as the worst poem in a child’s Mad Lib “fill out the rhyme” exercise.
I just asked Dall-E to make these images:
I’m not making fun of the AI. I’m pointing out that a human artist could quickly fulfill my request. So can a regular digital program. There have been DIY Magic 8 Ball generators online for years.
Creating the 8 Ball’s sides for someone else’s leisure and pleasure
The big AI companies are paying people pennies to “annotate” images so that the rest of humanity believes there is some sort of volitional intelligence behind AI art, chat, and large language models.
When I say “pennies,” I mean literal pennies.
In June, Josh Dzieza published an in-depth report in The Verge about workers who are training AI models. They’re called “Taskers” and the remote jobs have been available for years, but are rapidly accelerating as various AI development companies discover new needs for human intervention.
Dzieza interviewed one worker who said, “I read and I Googled and found I am working for a 25-year-old billionaire,” said one worker. “I really am wasting my life here if I made somebody a billionaire and I’m earning a couple of bucks a week.”
Dzieza also spoke with Erik Duhaime, the CEO of a medical data annotation company. At first, Duhaime said that ML engineers thought that AI would make the job of radiologist — a healthcare professional who interprets XRays and MRIs — obsolete.
Next, Duhaime said, developers thought that radiologists would use AI as a tool.
However, now that AI is in use, Duhaime said he doesn’t see that happening. According to Duhaime, in AI/ML-informed radiology, humans are now needed to:
Check that the AI is being fed the right type of data
Check the AI’s work before passing it on to another AI, which writes the imaging report
Check the imaging report
Duhaime said, “AI doesn’t replace work. But it does change how work is organized.”
So, in terms of radiology, my husband just had complete MRIs of his spine and other joints. The radiologist viewed and interpreted the results, writing a report. The neurosurgeons have also viewed and interpreted the results. They don’t write reports but their staff codes their desired recommendations for surgery. The codes are then sent to the insurance company. The insurance company has approved the codes for the surgeon’s request.
Later on, AI will send us a bill and we will need to take out a home equity loan to pay for the surgery.
Maybe not yet, but soon — a robot will do the surgery while the surgeon watches.
“I really am wasting my life here if I made somebody a billionaire and I’m earning a couple of bucks a week”
AI Colonialism — more widespread than most of us know
I’ve previously written that AI has the potential to free humans up to become truly creative. If we are not required to work for someone else for the majority of our lives, what then, would most people choose to do?
The opposite is happening so far — little surprise — as the technology is being controlled by individuals who want to live in a world where billions work to benefit a few thousand others.
The basic idea of AI as it stands and is being developed is deeply flawed, much as our food systems have been hijacked by profit-makers who will not alter one step in an environmentally-devastating process, replete with animal cruelty and human exploitation. It’s obvious from the conduct of AI-training programs and companies, whether they are ScaleAI or any competitor, that they have no regard for anything other than earning profit for their owners and investors, as quickly as possible.
MIT Technology Review has covered the 21st century colonialization of AI tasking/training work. What this means is that young people in countries that most people in the U.S. view as “poor” are recruited as faceless, nameless cogs in the AI wheel: they are the ones who categorize images (mostly advertising) and listen to phone calls and do other repetitive tasks like identifying elbows in a picture of a crowd.
The companies that hire international workers to provide annotation, labeling, and annotation services include Scale AI (the owner is the billionaire who inspired the worker to say he was wasting his life making him ‘rich’), Appen, Hive Micro, and Spare5. Over a year ago, MIT Technology Review reported that over 200,000 Venezuelans were working for these companies doing “ghost work.”
I hadn’t heard of the term, “ghost work” before researching this type of work before. I see that all of the usual better-paid, more socially powerful individuals than I am have not only already heard of it, they are already participating in the Nickeled and Dimed-style academic launch of a book by anthropologist Mary L. Gray.
Well, Gray et. al. and the students forced to read her book are just as trapped as the young man who wondered, “Am I wasting my life to make some 26 year old a$$hole a billionaire?”
This push has come and gone and it seems to have been just about as influential as Nickeled and Dimed, a book by Barbara Ehrenreich with a stomach-turning forward that I read right as I was getting kicked to the curb by my high-powered agent. Basically, Ehrenreich (who died in 2022), wrote about having a meal in Manhattan with her publisher (I don’t remember but maybe New Yorker or New York Magazine) where they arrived at the scheme by which she would work at minimum wage jobs and see how the people who had to do them lived.
Ehrenreich never lasted longer than two weeks at any job and she almost gleefully abandoned some of her co-workers in bad situations, especially during the “Merry Maids” chapter where she cleaned houses for about two weeks for $6.25/hour.
So, the basic situation is this. People go into work arrangements because they need money to pay for food, a roof over their heads, and clothing.
The people who “provide the work” often do not care if the individual doing the work lives or dies. They believe that another person will quickly take their place if they do die, or if they actually manage to get a better job and improve their lives.
I am writing this because a) I was recruited to teach one of these machines how to write. I already know that is impossible; and
b) unlike Ehrenreich, Gray, and the dozens of academics and influencers who recommend her “ghost work” book that states Silicon Valley companies like Scale AI are exploiting poor people in distressed nations to work for pennies to convince others that the machines are somehow “thinking” on their own and doing valuable services instead of sending me the same pictures of weird, imitation clothing from Temu over and over and over — (like that sentence? keep it up Pro Writing Aid, Grammarly, Tasker-influenced machine!) —
I’ve worked plenty of shit jobs and I know what it feels like to be hungry, frightened, lose my house and be forced to file for bankruptcy —
The answer really does lie within all of us. Will AI free people to conduct better, healthier, happier lives?
Yes: and the day that starts will be the day this young man and all the many other people like him decide they no longer want to waste their lives making a 26 year old a mega-billionaire by listening to people categorizing the emotions they hear in someone ordering pizza.
That’s what the guy was being paid pennies per call to do.
This morning, Business Insider reported that Amazon has built, or will be building, over 100 data centers in Northern Virginia, which together will use more power than a large city (New York, Chicago — you name it). And this type of worker, all over the world, is the primary reason that such data power is “needed.”
To create more answers by AI to questions that few, to no people, have ever asked.
To create dull, unoriginal pictures that are “good enough”.
To write dull, unoriginal poems that no one cares to read.
To write papers for students that neither they, nor instructors, read.
But in the end, it is just to serve ads. To make more billionaires and to make the billionaires we already have, even richer.
Can we not take Nancy Reagan’s advice and “Just say no?”
As a noted AI researcher said: "AI is neither."
I never understood the cachet Ehrenreich earned by slumming as a member of the permanent underclass. I remember reading Nickeled and Dimed, and while it was captivating on its surface, it more reeked of that liberal idea of charity: less totalitarian than the right, but ultimately only an incompetently softened gloss of the right's worst attitudes and practices. Ehrenreich didn't think she was doing the service she thought she was. Liberals still crow about her book, probably because it's the only experience any of them have with a $hit job.
AI is just the latest scam for our scarecrow economy that produces nothing of actual value, while throwing the word "value" around.