Executive Summary
The U.S. artificial intelligence workforce, which stood at 14 million people in 2019, or 9 percent of total U.S. employment, has grown rapidly in recent years. This trend is likely to continue, as AI occupational employment over the next decade is projected to grow twice as fast as employment in all occupations.
Such an important and increasing component of the U.S. workforce demands dedicated education and workforce policy. Yet one does not exist. To date, U.S. policy has been a piecemeal approach based on inconsistent definitions of the AI workforce. For some, current policy is focused on top-tier doctorates and immigration reform. For others, the conversation quickly reverts to STEM education.
This report addresses the need for a clearly defined AI education and workforce policy by providing recommendations designed to grow, sustain, and diversify the domestic AI workforce. We use a comprehensive definition of the AI workforce—technical and nontechnical occupations—and provide data-driven policy goals.
Our policy goals and recommendations build off of previous CSET research along with new research findings presented here. Previous research in this series defined the AI workforce, described and characterized these workers, and assessed the relevant labor market dynamics. For example, we found that the demand for computer and information research scientists appears to be higher than the current supply, while for software developers and data scientists, evidence of a supply-demand gap is mixed.
To understand the current state of U.S. AI education for this report, we manually compiled an “AI Education Catalog” of curriculum offerings, summer camps, after-school programs, contests and challenges, scholarships, and related federal initiatives. To assess the current landscape of employer demand and hiring experiences, we also interviewed select companies engaged in AI activities.
Our research implies that U.S. AI education and workforce policy should have three goals: (1) increase the supply of domestic AI doctorates, (2) sustain and diversify technical talent pipelines, and (3) facilitate general AI literacy through K-12 AI education.
To achieve these goals, we propose a set of recommendations designed to leverage federal resources within the realities of the U.S. education and training system. Our first recommendation sets the foundation for facilitating these goals by creating a federal coordination function. We believe such a function is critical given ongoing fragmented AI education initiatives, and would harness the potential of the newly established National Artificial Intelligence Initiative Office for Education and Training within the White House Office of Science and Technology Policy. We recommend this office coordinate federal and state initiatives, convene key stakeholders to share lessons learned and best practices of state-level AI education initiatives, and compile and publish information on AI education and careers on a publicly available “AI dashboard.”
The remaining recommendations advocate for a multipronged approach to implement policies across goals, including:
- Creating and Disseminating AI Educational and Career Information
- Establishing AI Education and Training Tax Credits
- Investing in Postsecondary AI Education and Scholarships
- Facilitating Alternative Pathways into AI Jobs
- Investing in PreK-12 AI Education and Experiences
- Integrating K-12 AI Curriculum and Course Design
- Cultivating and Supporting K-12 AI Educators
- Funding AI Education and Careers Research
Importantly, our recommendations prioritize creating multiple viable pathways into AI jobs to diversify the AI workforce and leverage all U.S. talent. Our research shows the dominant pathway to enter the AI workforce remains having a four-year college degree. However, this may be restricting the amount of talent entering the AI workforce, unnecessarily limiting opportunity for those who are otherwise qualified and able.
Our recommendations therefore prioritize harnessing the potential of community and technical colleges, minority-serving institutions, and historically Black colleges and universities in training tomorrow’s U.S. AI workforce. In addition, to promote alternative pathways into AI jobs, we propose that the National Institute of Standards and Technology work with industry to establish industry-accepted standards for AI and AI-related certifications to enhance their legitimacy. And as a top employer of technical talent, the federal government could modify its hiring criteria to lead by example.
We hope that this report and recommendations advance the discourse on AI education and workforce policy. Now is a critical time to invest in training and equipping a globally competitive AI workforce for tomorrow. With concerted and targeted efforts, it is possible to lead the world in AI talent. Ultimately, an AI workforce policy inclusive of all of our report’s elements is more likely to be the most effective. However, we also present our recommendations as a road map to guide U.S. policymakers in crafting an AI education and workforce agenda.