What are digital sweatshops?
"Digital sweatshops" is a term used to describe work environments where employees, often in developing countries, perform repetitive and low-paid tasks using digital technologies under poor working conditions.
Training AI models require manually labeled data. As a result, companies like Scale AI exploit foreign workers to label data under poor conditions: a digital workshop.
Why should I care?
How much are workers paid?
But at least they are paid, right?
What else should I know about Scale AI?
Digital sweatshops are a human rights issue. As the AI industry grows and begins to permeate all aspects of society, it's in our best interest to hold companies like Scale AI accountable for their malpractices so that AI developers understand that their choices are consequential and seen.​
​
Unfortunately, 7 billion dollar startup Scale AI, chooses to severely underpay workers and does not adhere to basic labor standards for outsourced work. Some workers report receiving significantly reduced pay.
​
One employee reported that four hours of work for an expected $2 payment resulted in only $0.30. Another accumulated over $150 in expected pay but was suddenly released without compensation.
​
To learn more, read this Washington Post article.
Sadly, some workers report that payments are regularly withheld. These workers cannot pursue missing compensation for fear of termination.
Others report being terminated from their position before receiving their pay, leaving them with nothing.
As such, Scale AI can exploit these workers without consequence.
Scale AI lacks transparency. They will not release information on the number of employees involved in data labeling. Talking at Stanford University, Scale AI threatened to leave if any students mentioned digital sweatshops.​