The Impact of Artificial Intelligence on Vulnerable Populations in the Workforce

To understand how AI can be integrated into the office, we need to first define the current workplace landscape. Independent of advances in AI, the workplace has undergone several transformations in the past decade. First, the pandemic caused major shifts in the workforce such as reduced customer-facing roles, a decline in food service, customer service and sales positions, and growth in other professions like management and health care. Second, the transition to a net-zero economy urges workers in the fossil fuel supply chain to reskill or retrain for the move to green industries. Finally, general trends of technology advance towards a gig economy, as seen in the development of app-based, flexible and contingent work jobs like ride share or food delivery services. Each of these trends will continue to develop not only parallel to AI, but also in reaction to AI. 

Generative AI can be defined as a set of tools that can “identify patterns across enormous sets of data and generate new content—an ability that has often been considered uniquely human.” It is this ability to venture into a space “uniquely human” that enables AI to impact a broader range of industries not impacted by automation alone. In considering how AI will affect the workplace overall, this article focuses on how AI acceleration and augmentation affects existing disparities in the job market. 

The Concerning Impacts of AI on Vulnerable Populations
Existing AI products have been found to mirror society’s biases, unless the products intentionally address bias throughout product design and deployment. Similarly, and specifically, the use of AI in the workforce has the potential to exacerbate existing race and gender disparities, unless the current development of AI products accounts for and protects against the harmful effects of AI in the workplace in the future.

AI is expected to impact a wide range of jobs, from office support, repetitive-motion jobs, customer service and sales positions, to mathematics and architecture. While both high-wage and low-wage jobs will be affected by AI, low-wage jobs are more susceptible to replacement, as in the following examples:

• Black and Hispanic workers are overrepresented in the 30 occupations with the highest exposure to automation and underrepresented in the 30 occupations with the lowest exposure;
• Compared to 58% of roles held by men, 79% of roles held by women could be affected;
• Roughly two-thirds of current jobs are exposed to some degree of AI automation and a quarter of current work tasks could be substituted by AI; and
• Low-wage earners, those earning $38,200 a year or less, are 14 times more likely to need to change occupations by 2030.

The deployment of AI is expected to boost productivity, produce efficiencies and create more high-wage jobs; however, there are no guarantees low-wage workers displaced by AI will be able to access the newly created opportunities. Reports show employers’ use of AI will lead to an increased need for STEM students. Women, Black and Latino workers are underrepresented in STEM professions and, unless STEM employers, governments and academic institutions work to address this discrepancy, AI will further exacerbate the trend. Upskilling and training become central components to ensure current barriers to employment are not further exploited. 

Protecting Employees and Setting the Groundwork to Benefit from AI
One of the challenges and opportunities of the AI transition is to protect vulnerable populations through upskilling and training programs. Employers and governments need to work together to prepare the workforce for the AI transition and assist workers to be competitive when new, higher-wage jobs are created. While AI is new, government programming to address AI does not need to reinvent the wheel. Legislative history and public-private partnerships, along with private sector precedent, can be leveraged to prevent the inequitable deployment of AI. 

Existing Federal Programming
The Inflation Reduction Act (IRA) and the Infrastructure Investment and Jobs Act (IIJA) both provide incentives for bolstering the workforce and may serve as templates for future congressional incentives addressing the AI transition. The IRA provides apprenticeship and prevailing wage tax credits encouraging companies to train new and retain existing employees. Several IIJA funding opportunity programs are actively encouraging partnerships with local educational institutions and community workforce leaders. The Department of Labor has a history of allocating funds to grant programs targeting industries or vulnerable populations, such as the Workforce Opportunities for Rural Communities grant which provides demonstration grants for the Appalachian, Delta and Northern Border regions. These examples serve as a guide as to how the government can invest in workforce development specific to AI—by creating AI-focused apprentice programs, encouraging workforce retention for companies affected by AI, and investing in educational institutions providing upskilling and training for low-wage workers to compete for the new high-wage jobs. 

Continued and Future Investment in Federal Workforce Initiatives
On October 30, President Biden signed the Executive Order on Safe, Secure, and Trustworthy Artificial Intelligence, which directs federal agencies to responsibly deploy AI, including by mitigating risks to the workforce. The Executive Order provides that “all workers need a seat at the table” to ensure that all workers benefit from the opportunities AI presents. Federal Agencies will take the following actions to pave the way for the AI workforce transition:

• The Council of Economic Advisors will Report on the labor-market effects of AI;
• The Department of Labor will report on the role of federal agencies in supporting workers;
• The Department of Labor will create best practices and standards to protect employees; and
• The National Science Foundation will prioritize AI-related education in existing programs.

Through the Executive Order, the Secretary of Labor is directed to create two reports that will be important to safely navigate the AI transition. First, within 180 days of the Executive Order, the Secretary of Labor will create a report reviewing the ability of federal agencies to support workers displaced by AI, in order to evaluate further steps the federal government can take to address workforce disruptions. This report will also review how existing programs can be leveraged to address AI impacts on the labor market. The report may provide policy and legislative suggestions to augment federal support for workers as well as increase education and training opportunities related to AI occupations.

Second, also within 180 days of the Executive Order, the Secretary of Labor will collaborate with other agencies, unions and other outside entities to develop a set of best practices that employers could use to mitigate AI risk to employees. The best practices are expected to cover career risks and opportunities, issues relating to labor standards and job quality, including how AI might affect equity in the workplace, safety and health, and how employers use AI and AI-related data collection regarding their employees. Guidance provided by the Secretary of Labor will make clear that AI used by employers in the workplace regarding their employees must abide by existing labor laws. Together, the activities from the Department of Labor will outline a path forward for federal government and private-sector employers to protect employees from AI while maximizing its potential benefits.

Public-Private Partnerships
The government may work in concert with industry in carrying out workforce programming. Engagement between the government, the private sector and community organizers can address the AI ecosystem wholistically. One report encourages organizations to work across communities to “provide training, upskilling and lifelong learning for workers, especially as current regulations in some geographies, including in the U.S., preclude organizations from providing training to contractual workers.” Access to educational opportunities, such as digital learning, will broaden accessibility and ease for displaced workers to quickly recover, receive training and find new employment.

Addressing upskilling and training through public-private partnerships supported by federal dollars can support employees affected by AI in a wholistic manner. Pillsbury’s attorneys have extensive experience navigating the appropriations processes, enabling them to identify opportunities to integrate AI workforce protections into existing programs and capitalize on opportunities to fund new programs. 

Conclusion 
AI is expected to boost national GDP and economic productivity, with some estimating the adoption of AI in the workplace increasing productivity by an impressive 1.5 percentage points. AI holds the potential to deliver lower prices, create efficiencies and provide convenience across the globe. However, unless AI is deployed intentionally to account for existing workplace disparities, it has the potential to raise existing high barriers to access for populations. With input from communities and increased access to upskilling and training opportunities, it is possible to create an environment allowing all people, regardless of gender or race, to gain from the growth spurred by AI. 

Pillsbury stands at the ready to help companies prioritize equity for their employees. Pillsbury can help companies be proactive in mitigating potential AI harms through initiatives like upskilling and training programs and help companies prepare to benefit from AI progress. Pillsbury’s multidisciplinary team of AI thought leaders and legal and strategic advisors is an industry leader in strategic promotion of responsible and beneficial AI. Pillsbury closely monitors AI-related legislative and regulatory efforts, helping startups, global corporations and government agencies navigate the landscape impacted by emerging developments in the technology.  For insights on these rapidly evolving topics, please visit our Artificial Intelligence practice page.