9 Comments
User's avatar
Celeste Garcia's avatar

It's by design. Big Tech does not want you to know about the hundreds of thousands of human laborers required to refine the data that goes into AI models like Llama, ChatGPT, Copilot, and Claude. This could be an opportunity to offer well-paying jobs in developing countries if technology companies were willing to provide fair wages and good working conditions. Many of these workers make less than $10 for 8 hours of tedious, often emotionally disturbing work, if they get paid at all. This is happening simultaneously to Zuckerberg offering nine-figure compensation packages to individual "AI prodigies," reportedly ranging from $100 million to $1.5 billion over several years to join his Superintelligence Labs. It's abhorrent.

Expand full comment
Celeste Garcia's avatar

Yep, super glad you pointed this out, Michael. Founders Fund was an early backer of Scale AI & was influential in the $100M round in 2019. Thiel has nothing but praise for Scale AI’s role in the AI ecosystem. Thiel looks for founders he could “become good friends with.” This speaks volumes about the character of Wang.

There are so many layers to this story. My 1st post was an attempt to explain the deeply complicated labor ecosystem Wang built. My next installment will focus on why the DOL abruptly dropped the investigation on May 9, 2025, and requires a lot of research & digging. It’s no secret that Wang has been courting Donald J, plus Michael Kratsios, our new Director of Science & Tech to the White House was Managing Director at Scale AI after serving as CTO for the 1st Trump administration.

Expand full comment
Shan's avatar

Thanks for giving voice to exploited labor. I had no clue so many humans are need for the the tasks you mention.

Expand full comment
Tajana Ida's avatar

It is a great observation, Celeste and urgent to point out all the time and everywhere: "Turns out AI isn’t magic. Armies of flesh-and-blood laborers are needed to label, tag, correct errors, and sort raw data that is fed into our beloved AI models like ChatGPT, Claude, and Llama. This data refinement is performed by a vast global workforce, including hundreds of thousands of gig workers."— I would add "... laborers..., as well as underpaid workers skilled in their areas of expertise, for the corrections of substantial errors..."

Expand full comment
Ryan Sears, PharmD's avatar

Absolutely despicable business practices. Disappointed but not surprised to hear all the big names are also complicit.

Thank you for sharing this very important research, Celeste.

Expand full comment
Michael Spencer's avatar

You don't dig deep enough, this is another Peter Thiel plant. Scale AI wasn't even Mark Zuckerberg's first choice. But their lack of a moral compass is certainly aligned. But this is all by design, certain key Venture capitalists and Tech executives have a very defined agenda for the future of AI.

Expand full comment
Flavia Sekles's avatar

Well written and well researched, Celeste. So glad that you are -- indeed -- a writer. It strikes me how historically repetitive this story is.

Expand full comment
Rosana Francescato's avatar

Yikes! This is really shocking, and even more shocking (though no longer surprising) is legacy media's ignoring of this story. So disturbing. Thanks for bringing this to light.

Expand full comment
Celeste Garcia's avatar

Indeed. History repeating itself. It's so sad that we haven't progressed beyond the exploitation of impoverished people in underdeveloped countries by the richest humans and companies on the planet--Scale AI, Meta, Amazon, Google, Microsoft, Apple, xAI, Anthropic, OpenAI. With your UNLIMITED resources, it's not that hard. Do the right thing.

Expand full comment