Labor Dept. outlines worker-centric approach to AI in new roadmap

The Department of Labor has issued new principles tailored to workplaces utilizing artificial intelligence that focus on worker wellbeing. 

The agency’s Artificial Intelligence Best Practices roadmap details steps workplaces and software developers can take to ensure their use of AI doesn’t disempower employees. As voluntary guidance, the document sets no legal obligations mandating developers and employers using AI follow Labor’s guidelines.

Labor’s Acting Secretary Julia Su noted that as AI tools become more ubiquitous across diverse industries, its primary objective should be to aid and expand workers’ opportunities. 

“I want to be really clear about one thing: The false narrative that we have heard that we have to choose between innovation on the one hand and worker wellbeing on the other is just that — it's a false narrative,” she said during a virtual meeting Wednesday. “Workers have been the fuel for innovation in this country and around the world, and what we have to ensure is that the innovation benefits workers, it enhances what human ingenuity and ability can do, and it makes life better for working people, who are the backbone of any society.”

Su said that AI in the workplace shouldn’t be used in discriminatory or retaliatory tactics that deepen existing inequity. The “North Star” for Labor is worker empowerment, and is reflected in the other seven recommended principles: ethically deploying AI; establishing governance and human oversight; ensuring transparency in AI deployment; protecting labor rights; using AI to enable workers; supporting workers impacted by AI; and ensuring responsible use of worker data.

Labor union AFL-CIO approves the new recommendations, with its president, Liz Shuler,noting the particular importance of human oversight. She added that American workers are broadly still concerned about AI’s presence in the workplace and society, citing a 2023 poll that suggests seven out of 10 workers are worried about AI threatening their employment.

“It's pretty clear when we look around us, we are in the wild west right now,” Shuler said. “These technologies can be really disruptive and have zero guardrails, but there's also a massive opportunity, if we get this right, to use these technologies in a way that benefits everyone.”

In keeping with the theme of a worker- and human-centric approach to AI, the roadmap noted that the early and frequent integration of feedback and participation from workers in the adoption process of AI systems will improve both the workers’ job quality and the workplace’s desired outcome with AI. 

“The decisions that we make today are going to shape the impact of AI on workers for years to come,” Su said.