From the Philippines to Colombia, low-paid workers label training data for AI models used by the likes of Amazon, Facebook, Google, and Microsoft.

“The global data collection and labeling market was valued at $2.22 billion in 2022 and is expected to grow to $17.1 billion by 2030, according to consulting firm Grand View Research. As Venezuela slid into an economic catastrophe, many college-educated Venezuelans like Fuentes and her friends joined crowdsourcing platforms like Appen.”

  • bane_killgrind
    15•9 months ago

    These jobs are the actual valuable work that is being done around AI. There’s a huge problem of poisoned and dirty data sets, evidenced by chatbot Nazi diatribes and otherwise garbage outputs.

    That being said, these sweatshop apps are fractionally deriving value from nothing but facilitating communication and branding, things that are very cheap to implement. As more players enter the market, that paradigm will either get worse or better for the workers that are trying to live their lives.

    Corporations will divorce workers and the communities they live in from the value of their labors.

    • @[email protected]English
      11•9 months ago

      I did this for awhile and a huge problem was that people were training the models using content generated by ChatGPT. Trash goes in, trash comes out. That’s the quality you get when you pay people peanuts I guess.