A Proposed AI Task Force to Confront Talent Shortage and Workforce Changes

Just over a month after House and Senate commerce committees received companion bills recommending a federal task force to globally examine the “FUTURE” of Artificial Intelligence in the United States (H.R. 4625; introduced Dec. 12, 2017), a House education and workforce committee is set to consider a bill calling for a task force assessment of the impacts of AI technologies on the US workforce.

If enacted, the “Artificial Intelligence Job Opportunities and Background Summary Act of 2018,” or the “AI JOBS Act of 2018” (H.R. 4829; introduced Jan. 18, 2018), would require the Secretary of Labor to report on impacts and growth of AI, industries and workers who may be most impacted by AI, expertise and education needed in an AI economy (compared to today), an identification of workers who will experience expanded career opportunities from AI and those who may be vulnerable to career displacement, and ways to alleviate workforce displacement and prepare a future AI workforce.

Assessing these issues now is critical. Former Senator Tom Daschle and David Beier, in a recent opinion published in The Hill, see a “dramatic set of changes” in the nature of work in America as AI technologies become more entrenched in the US economy. Citing a McKinsey’s Global Institute’s study of 800 occupations, Daschle and Beier conclude that AI technologies will not cause net job losses. Rather, job losses will likely be offset by job changes and gains in fields such as healthcare, infrastructure development, energy, and in fields that do not exist today. They cite Gartner Research estimates suggesting millions of new jobs will be created directly or indirectly as a result of the AI economy.

Already there are more AI-related jobs than high-skilled workers to fill them. One popular professional networking site currently lists over 6,000 “artificial intelligence” jobs. Chinese internet giant Tencent estimates there are only 300,000 AI experts worldwide (recent estimates by Toronto-based Element AI puts that figure at merely 90,000 AI experts). In testimony this week before a House Information Technology subcommittee, Intel’s CTO Amir Khosrowshahi said that, “Workers need to have the right skills to create AI technologies and right now we have too few workers to do the job.” Huge salaries for newly-minted computer science PhDs will drive more to the field, but job openings are likely to outpace available talent even as record numbers of students enroll in machine learning and related AI classes at top US universities.

If AI job gains shift workers disproportionately toward high-skilled jobs, the result may be continued job opportunity inequality. A 2016 study by Georgetown University’s Center on Education and the Workforce found that “out of the 11.6 million jobs created in the post-recession economy, 11.5 million went to workers with at least some college education.” The study authors found that, since 2008, graduate degree workers had the most job gains (83%), predominantly in high-skill occupations, and college graduates saw the next highest job gains (57%), also in high-skill jobs. The highest job growth was seen in management, healthcare, and computer and mathematical sciences. These same fields are prime for a future influx of highly-skilled AI workers.

The US is not alone in raising concerns about job and workforce changes in an AI economy. The UK Parliament’s Artificial Intelligence Committee, for example, is confronting challenges in re-educating UK’s workforce to improve skills needed to work alongside AI systems. The US may need to do more to catch up, according to Mr. Khosrowshahi. “Current federal funding levels [in tech education],” he argued, “are not keeping pace with the rest of the industrialized world.”

The AI JOBS Act of 2018 presents an opportunity for US policymakers to develop novel approaches to address expected workforce shifts caused by an AI economy. If nothing is done, the US could find itself at a competitive disadvantage with increasing economic inequality.

Leave a Reply

Your email address will not be published. Required fields are marked *