A recent study conducted by MIT and Stanford examined whether an AI assistant based on OpenAI’s text-generation technology would make workers more productive. They found the AI tool boosted the support team's productivity by 14 percent. 

The results also raised a new question: Should workers be compensated when their data helps train an AI system to do their job? 

If chatbots like the one tested in the MIT and Stanford study become common, some workers might use their own power to push for new approaches to compensation. Aaron Benanav, assistant professor of sociology and senior research associate in the Autonomous Systems Policy Institute, points to companies in countries with friendlier collective bargaining laws, like Germany and Sweden, which tend to invest more in their workers than corporations in the U.S.

Surveys indicate that Swedish citizens display less anxiety about robots taking their jobs, in part because when companies introduce new technologies, they often pay to upgrade their workers’ skills. “If you upskill workers, you pay them more,” Benanav says. “That's a more durable and sustainable process.”

Read more in the WIRED article, “You Trained the Chatbot to Do Your Job. Why Didn’t You Get Paid?”