Artificial intelligence has graced humanity with many blessings: videos of Stephen Hawking as professional wrestler, thousands of AI-generated podcasts, and more cutting edge Spotify tunes than anyone can hope to hear in a lifetime.
Unfortunately, the technology of the future demands a high price. On top of the exorbitant energy cost fueling a return to industrial-era levels of pollution, AI is also propped up by a massive global sweatshop operation, where low-wage workers in underdeveloped countries are tasked with doing the hidden intellectual labor that makes the tech useful.
As reported by Agence France-Presse, workers in long-exploited countries like Kenya, Colombia, and India are becoming increasingly outraged over the miserable labor of AI training. For example, as the wire service notes, for an AI chatbot to generate an autopsy report, contract workers have to sift through thousands of gruesome crime scene images, a gig known as “data labeling.”
Though the work is often done remotely — thus saving on the overhead costs of leasing an office — data labeling isn’t exactly a cushy laptop job. Workers involved in this industrial operation describe grueling hours, few if any workplace protections, and frequent tasks involving violent or grisly content. In theory, it’s not unlike social media content moderation, another digital practice built on exploitative labor in the developing world.
“You have to spend your whole day looking at dead bodies and crime scenes,” Ephantus Kanyugi, a Kenyan data label, told AFP. “Mental health support was not provided.”
Adding to the nastiness, these contract workers aren’t employed by AI companies like OpenAI or Google directly. Instead, those corporations partner with third-party contractors, who then hire out data labelers like Kanyugi, whose home country has no laws regulating data annotation work.
One of the most well-known firms trafficking in the AI industry is Scale AI, which often works through various subsidiaries and shell companies. Scale AI is arguably the largest company in the seedy data labeling business, boasting deep ties to Silicon Valley behemoths like OpenAI and Meta, as well as esteemed clients like the US Pentagon.
Scale AI’s own labor practices are horrid enough. But AFP reports that one Scale AI subsidiary, Remotasks, is taking it a step further: paying data labelers roughly one US cent for every task they finish, which can take hours. It’s a system Kanyugi compared to “modern slavery.”
“People develop eyesight problems, back problems, people go into anxiety and depression because you’re working 20 hours a day or six days a week,” he told AFP. “Then despite working so many hours, you only get poor pay, and you might also not get paid.”
So next time you fire up ChatGPT — or ask Grok “is this true?” — spare a thought for the worker who made your chatbot possible in the first place.
More on AI: Parasitic Startup Pollutes Job Market by Applying to Jobs for You Automatically
The post The AI Industry Is Traumatizing Desperate Contractors in the Developing World for Pennies appeared first on Futurism.
