The article suggests that tech companies should pay data workers more and improve their working conditions. It draws a parallel with Nike's response to backlash over poor working conditions in the 1990s, suggesting that similar pressure could lead to improved conditions for data workers. The article also warns that poor working conditions can lead to substandard AI, which is problematic given increasing public wariness of AI models.
Key takeaways:
- Between 150 million and 430 million people work in the AI supply chain, often in low-paying jobs that involve training algorithms, with little chance of upward mobility.
- Many data workers live below the poverty line and often have to take second jobs or night shifts. The job itself is precarious and wages are often below local minimum wage.
- End customers of these data workers are often some of the most valuable firms in the world, such as Microsoft Corp. and OpenAI, but the workers themselves often struggle to make a living.
- Researchers argue that poor working conditions and low pay for data workers can lead to substandard AI, and advocate for better pay and treatment for these workers.