China’s surge in the AI sector reveals a disturbing truth: its progress depends on overworked, underpaid vocational students, creating a new digital underclass, journalists Viola Zhou and Caiwei Chen report in Rest of World.
The article features a computer science student from a Shandong vocational school, Lucy, who is among hundreds of thousands pushed into data annotation roles by their schools as the students must complete internships during their formation. They tag images, screen videos, and filter audio, often for wages that barely cover daily expenses, despite their pivotal role in refining AI models and advancing technologies like autonomous vehicles.
This exploitation is rampant in collaborations between Chinese data labeling firms and vocational schools. Lured by the allure of AI skills and potential job prospects, students end up in monotonous, assembly-line roles. Investigations show many are paid below minimum wage and work in harsh conditions, even when it conflicts with China’s educational regulations.
Historically, China’s vocational schools supplied various industries with affordable labor. Today, AI companies place students from disciplines like “Computer Science” and “Big Data” at the lowest wage tier. Academic researcher Xia Bingqing compares many data annotation internships to manual labor, a sentiment echoed by students. China mirrors a global pattern where tech giants outsource data annotation to poorer countries.
These arrangements financially benefit data companies, reducing costs with student interns. A manager at a data firm disclosed to Rest of World that many of his annotators were vocational students, with schools often taking a slice of their minimal wages. Additionally, some students must deal with disturbing content, including violent and mature material, leading to trauma.
With youth unemployment rising, many students view data annotation as an avenue to bolster their resumes or make ends meet. However, Lucy’s experience speaks volumes: “It wasn’t even a stepping stone.”