Building Worker Power in the AI Age: A Department of Labor Blueprint

Building Worker Power in the AI Age: A Department of Labor Blueprint

with Seth Harris and Dane Gambrell

reposted from The Reboot Democracy The Governance Lab Blog in collaboration with the Power At Work

The U.S. Department of Labor has issued a groundbreaking blueprint to ensure workers—not just employers—shape how AI is used in the workplace. Released in October 2024, these recommendations guide employers and tech developers on implementing AI in ways that enhance rather than undermine worker power—from requiring worker input on AI decisions to protecting organizing rights to mandating transparency about AI systems. These guidelines matter because they provide the first comprehensive roadmap for ensuring AI serves as a tool for expanding worker power and improving workers’ lives and economic security.

While these guidelines establish important principles for AI implementation in the workplace, they remain voluntary recommendations rather than binding requirements. Their effectiveness will depend largely on employer adoption—and all employers should embrace them. Similarly, unions should use them when negotiating with employers over AI-related topics.

It’s worth noting that the Labor Department does not have jurisdiction over all aspects of employers’ use of new technologies. Nonetheless, with a mission focused on improving the condition of working people, the agency’s words carry substantial weight.?

Five key recommendations stand out as particularly crucial for building worker power in the AI era:

1. Elevating Worker Voice as the "North Star"

The recommendations call for involving workers in AI decision-making. The guidelines offer helpful advice about how to implement this "north star" principle, calling for employers to establish formal channels through which workers can influence how AI is used in their workplaces.

Employers are urged to seek "genuine input" from workers and their representatives at every phase: design, development, testing, training, deployment, and ongoing oversight of AI tools, enabling them to put their workplace knowledge to use helping their employers choose the best tools and develop the most effective implementation strategies to "determine how AI can further support worker productivity, performance, and well-being."?

For example, when it decided to roll out access to AI and AI training for its workforce, the State of New Jersey surveyed workers about their views on AI, asking what they wanted to learn.?

The guidelines place special emphasis on engaging diverse participants and voices, particularly workers from communities that have historically been excluded from workplace decision-making processes. This includes people of color, women, immigrants, veterans, individuals with disabilities, LGBTQI+ individuals, rural workers, workers without college degrees, and others who have faced systematic barriers to full and successful participation in the workplace. By ensuring these workers have meaningful input in how AI systems are designed and deployed – such as when discussing AI in hiring or in measuring workplace safety – employers can help prevent these technologies from perpetuating existing workplace disparities or creating new barriers to opportunity.

USDOL

2. Establishing Essential Transparency Rights

Knowledge is power or, at least, a precondition for meaningful worker engagement. The guidelines recommend that employers provide clear disclosures about use of AI systems, especially those that are used within the workplace as part of a productivity monitoring system. The disclosures should include:

  • Advance notice to employees before implementing new AI tools
  • Details about what worker data is being collected and why
  • Employee rights to view and correct personal data used in employment decisions (without a well-founded fear of retaliation, see below)
  • Plain-language explanations of how AI affects significant workplace decisions

Many governments like the City of Boston, for example, mandate disclosure of AI use to workers and citizens alike.?

This transparency is fundamental to worker power. Without knowing what AI systems are being used and how they work, workers and their unions cannot provide information, input and insights.?

This transparency requirement is especially powerful for unions, which have both legal rights and collective power to demand information that individual employees typically cannot access. These disclosure rights become particularly important when considering how AI might affect workers' organizing rights.

3. Protecting Organizing Rights in the AI Era

The guidelines take a strong stance on protecting workers' rights to engage in collective action around AI-related issues. They plainly and clearly declare that employers should not use AI systems to bust or frustrate union organizing efforts. DOL also urges employers to notify workers and their representatives about any technologies that could potentially monitor organizing activities.

These protections are increasingly crucial as tech companies continue to build ever more sophisticated, powerful and intrusive workplace surveillance software. Reportedly, Amazon is already using AI tools to monitor workers both on and off the job in an attempt to identify and disrupt union organizing in their facilities.?

In 2022, the National Labor Relations Board General Counsel announced plans to prevent employers from using AI to monitor workers seeking to unionize.

The guidelines also affirmed that it is illegal to retaliate against workers who raise AI-related concerns, whether individually or collectively. This protection creates space for workers to speak up about employers' misuse of AI without fear of punishment.

4. Strengthening Collective Bargaining Over AI

If employees are represented by a union, these guidelines advise employers to negotiate in good faith with the union before implementing electronic monitoring systems or other types of AI tools in the workplace.?

We're already seeing unions tackle AI's role and use at the bargaining table. Following the strikes that shut down the entertainment industry for several months last year, the Writers Guild Of America, USA (WGA), the International Alliance of Theatrical Stage Employees IATSE and the @screen-actors-guild(SAG-AFTRA) each reached contracts with Hollywood studios that established AI guardrails designed to protect workers against displacement and worsened pay, without trying to banish AI from the entertainment industry altogether. While not without flaws , the Hollywood deals show how collective bargaining can serve as a tool to establish rules and practices for AI use that serve the interests of workers, employers, and developers alike.?

AI is not unique in this regard. Technology-driven job displacement has also been a centerpiece of the negotiations between the International Longshoremen's Association and the operators of the East Coast and Gulf Coast ports. More than 45,000 dock workers went on strike in early October, in part due to concerns that automation and advances in technology would lead to workers being replaced by machines. While the two parties reached a tentative agreement to boost wages that effectively ended the strike for the time being, they continue to negotiate about how to address the issue of automation.?

5. Building Power Through AI Skills Training

The guidelines emphasize that employers should take the lead in helping workers develop the skills needed to work effectively with AI systems, including by:

  • Providing training opportunities to help workers use new AI tools
  • Prioritizing retraining and internal reallocation when AI displaces workers
  • Partnering with state and local workforce systems to support education and upskilling

While countries like India and Finland are teaching AI to their entire population, with so many different kinds of AI it’s imperative that companies train their employees in the tools they use.

AI skills training builds worker power in two crucial ways. First, developing an understanding of AI and its capabilities enhances individual worker power. As more roles incorporate AI and automation, workers with AI skills are better positioned to retain their jobs and advance their careers. These skills also increase mobility, giving workers more options to change roles in search of better conditions and compensation.

Second, AI skills strengthen collective power. Understanding how AI systems work helps workers and their representatives recognize, challenge and bargain over issues such as job responsibilities and roles, career ladders, and training opportunities, as well as the possible unfair application of AI tools at work, including invasive monitoring practices.?

From Recommendations to Action

Strong unions will be essential to ensure that these guidelines translate into real protections and benefits for workers. Through collective bargaining, unions can negotiate specific terms about when and how AI will be used, establish clear limits on automated decision-making and worker surveillance, secure commitments for worker training and advancement opportunities, and ensure that productivity gains from AI lead to better wages and working conditions. Without strong union representation, workers may lack the collective power needed to make these guidelines meaningful in their daily work lives.

Employers’ adoption of AI for some applications is inevitable. The critical question is whether it will enhance or undermine worker power and improve or degrade workers’ lives and economic security. The Labor Department’s guidelines provide a roadmap for ensuring AI becomes a force for shared progress rather than deeper inequality. Workers, unions, and forward-thinking employers must now seize this moment to turn these principles into workplace reality.


Faisal J.

Experienced communications leader and award-winning broadcast journalist, utilizing years of learnings on both sides of the story to give unparalleled guidance and execution on strategic and crisis communications.

2 周

Interesting

回复
Faisal J.

Experienced communications leader and award-winning broadcast journalist, utilizing years of learnings on both sides of the story to give unparalleled guidance and execution on strategic and crisis communications.

2 周

Intresting

回复
Natalie Evans Harris

MD State Chief Data Officer | Expert Advisor & Speaker on responsible data use | Leading initiatives to combat economic and social injustice with the Obama & Biden Administrations, and Bloomberg Philanthropies.

3 周

This is a crucial conversation as we navigate the AI landscape. It's vital that all stakeholders collaborate to ensure that technology uplifts workers instead of leaving them behind. Together, we can create a future where everyone benefits! Beth Simone Noveck

Alexander Howard

Dad, writer, cyclist, citizen, cereal dilettante. Earnest advocate for freedom of information, open governance, & digital democracy. Lover of the Oxford Comma, hater of hubris. Trying to move carefully & fix things.

3 周

I’d bet on both.

回复

The recommendations contained in the "roadmap" are now available in StratML Part 1, Strategic Plan (ISO 17469-1) format at https://stratml.us/drybridge/index.htm#AIWWB or, more specifically, https://stratml.us/docs/AIWWB.xml Here's ChatGPT's conclusion about the potential to leverage the StratML standard and apply AI to ensure that AI applications empower rather than replace workers: "StratML Part 2 provides a robust, machine-readable format that facilitates transparency, accountability, and continuous improvement in AI model performance. Adopting it for AI performance plans would standardize metrics, clarify stakeholder roles, and streamline reporting, ultimately driving more ethical and effective AI development. This approach would ensure that all stakeholders—from developers to end-users—have a clear understanding of model performance, roles, and responsibilities, fostering trust and responsible AI deployment." https://chatgpt.com/share/6723b3e8-c3c0-800b-a3e8-28a17d789d99 BTW, it would be good if DOL were to demonstrate leadership by example in complying with section 10 of the GPRA Modernization Act.

要查看或添加评论,请登录

社区洞察

其他会员也浏览了