That AI You’re Using Was Trained By Slave Labor, Basically

The article discusses the exploitation of workers in the AI industry, particularly those involved in AI-labeling tasks. These workers, often from countries with low wages such as Venezuela, the Philippines, Lebanon, and Kenya, are paid significantly less than the minimum wage in their respective countries. They work long hours, often from their homes, tagging algorithm training data for companies like Microsoft, Google, and Amazon. The tasks pay between two and 50 cents each, and workers sometimes earn only a dollar or two per day.

The article highlights the experiences of two such workers, a Venezuelan national living in Colombia and a teenager from Pakistan, both of whom describe their work as "digital slavery." Despite the low wages and long hours, they continue to work due to the lack of other job opportunities. The article criticizes companies like Appen, Clickworker, and Scale AI for outsourcing their labor at such low wages, comparing the situation to the state of American prison labor.

Key takeaways:

AI-labeling workers, often from low-wage countries, are paid very little for their work, sometimes less than the minimum wage in their countries.
These workers often work long hours, completing 'microtasks' for companies like Microsoft, Google, and Amazon, and can earn as little as one dollar per day.
One worker described the situation as 'digital slavery', due to the low pay and long hours required.
Companies like Appen, who employ these workers, are reportedly trying to improve conditions, but there is a balance to be struck between the speed of task completion and the consistency of work for their laborers.

That AI You’re Using Was Trained By Slave Labor, Basically

Key takeaways:

Comments (0)

Newsletter