Department of Labor Issues AI Best Practices

On October 16, 2024, the Department of Labor (DOL) issued its first guidance on the use of artificial intelligence (AI) in the workplace. These AI Principles and Best Practices are not binding law or regulation, but instead offer employers a guide to avoid violating long-standing laws when using AI. This document was developed in response to President Biden’s Executive Order calling for the safe, secure, and trustworthy development and use of AI. The goal of the DOL’s publication is to mitigate the harm AI could do to worker well-being while maximizing its benefits.

These best practices are guided by eight principles. The first principle is what the Department is calling the “North Star” for the use of AI: centering worker empowerment. It calls for workers to be active participants in the development of any AI that an employer is going to use in the workplace. This includes input on the design, testing, training, use, and oversight of the AI systems they will be using.

The guidelines also call for ethically developing the AI as well as establishing strong governance and human oversight of its use. In sum, employers should design, develop, and use AI in a manner that protects workers. The process should not stop once the AI is implemented. Employers should constantly assess and audit their AI to ensure it continues to achieve maximum efficiency with minimal worker harm. Additionally, companies should have clear systems of AI governance. These systems will dictate when to use AI and the limits of how it is used. Also, the Department seeks to protect workers by advising companies to ensure humans hold the ultimate oversight over workplace AI. In other words, AI should not be watching over itself; someone needs to make sure it is not harming employees. For example, AI should not be permitted to make hiring decisions; it can make recommendations, but a human ultimately needs to decide who to hire. Going hand-in-hand with all of this is the Department of Labor’s recommendation that employers are transparent about their AI use. Companies should be clear with both workers and applicants about the use and role of AI in the workplace. They should also consistently publish the results of their continued assessments of the use of AI.

The last group of guiding principles seeks to actively protect workers. Employers should ensure that the use of AI will not infringe on their workers’ rights. This includes rights to organize, the right to a safe workplace, wage and hour rights, and anti-discrimination/anti-retaliation rights. Furthermore, AI should be used to assist workers, not replace them. As part of this, companies should look to support those impacted by AI. This includes training and transition assistance for those directly impacted by the introduction of AI into the workplace. Lastly, the Department of Labor strongly urges employers to ensure that AI responsibly uses worker data. Companies seeking to implement workplace AI should only use worker data that is necessary to support the legitimate business goals of the AI introduction. And the data collected should be treated with the requisite confidentiality and protections as any other sensitive data.

In general, the Department has taken a very worker-friendly approach to the implementation of AI in the workplace. While this document is voluntary guidance for now, it could be a preview of where regulation in this space will ultimately land. This foresight could give employers a head start should DOL introduce rules and regulations on workplace AI.