The State of AI-Related Apprenticeships

Executive Summary Over the past decade, the creation and adoption of increasingly sophisticated artificialintelligence systemshavesparked a new demand for middle-skill AI talent;that is,workerswho neither hold a bachelor’s nor an advanced degree.During that same timeperiod, policymakers have committed increased attention and resources torevitalizingthe nation’s Registered Apprenticeship system. The rising demand for middle-skill AItalent, combined with increased interest in a workforce development approach that hashistorically targeted middle-skill occupations, is the impetus for this brief examining thetopic of AI apprenticeships. The apprenticeships we describe in this brief are in occupations that share some of theknowledge, skills, and abilities (KSAs) with the range of competencies needed to createAI systems. As such, the apprentices highlighted are not necessarily currently workingin AI-specific roles. However, they do possess KSAs that would allow them toparticipate in the design, development, or deployment of an AI product. This approachallows us to measure the amount of potential talent that could work in the AIecosystem. We refer to apprenticeships in such occupations as AI-relatedapprenticeships. In our analysis, we assessedtrends in the number, completion rates, demographics,geographic distributions, and program sponsors of AI-related apprenticeships between2013 and 2023,using a novel dataset that draws on the Department of Labor’spublicly available Registered Apprenticeship Partners Information Database System(RAPIDS). We identified a total of over 19,000 new apprentices in AI-relatedoccupations. As a whole, our findings show that over the past decade, AI-related apprenticeshipshave become a more common training pathway for the technology workforce. Theseprograms have high completion rates, can reach underserved populations, aregeographically dispersed, and are used by small-to medium-sized firms to recruitsmaller numbers of workers. ●AI-related apprenticeships were practically nonexistent in2013 buthavesince expanded rapidly.The first three years of our data, 2013 through 2015,had very few apprentices in AI-related occupations. Since 2015, programs haveregistered 18,980 new apprentices. During the highest period of growth, 2020to 2022, the number of new apprentices in AI-related occupations increased by191%, much higher than the rate of growth for all apprenticeships. Center for Security and Emerging Technology |1 ●AI-related apprenticeships have extremely high completion rates.On average,68% of apprentices in AI-related occupations completed their program, which is25 percentage points higher than the completion rate of all (non-military)apprenticeships. ●The number of organizations sponsoring AI-related apprenticeship programsalso increased over time alongside the number of new apprentices.A smallnumber of sponsors account for a large proportion of registered apprentices inAI-related occupations, but there are also hundreds of sponsors that haveregistered fewer than ten apprentices. ●AI-related apprenticeships have effectively recruited Black apprentices.Thepercentage of Black apprentices present in AI-related occupations far exceedsthe percentageof Black workers in AI and STEM fields. In 2022, the number ofnew Black apprentices in these fields nearly equaled the number of new Whiteapprentices. ●AI-related apprenticeships are less effective at reaching women and Hispanicapprentices.Gender disparities present in other STEM fields and in the broaderAI workforce are also reflected in AI-related apprenticeships. Meanwhile,Hispanic and Latino representation in AI-related apprenticeships is lower (12%)across all years compared to theiroverall participation in apprenticeships (20%)from 2015 to 2024. ●AI-related apprenticeship programs are located in nearly every state, but thetop ten states by new apprentices registered account for a huge share of AI-related apprenticeships.From 2013 through 2023, the top ten performingstates registered 62% of new apprentices in AI-related occupations. Five ofthose states are not among the most populous states. Missouri, Texas, andCalifornia were the top three states in terms of new AI-related apprenticesregistered. To sustain progress, federal and state governments should continue to supportapprenticeship initiatives, thereby solidifying apprenticeships as a valuable pathway forworkers in AI-related and other technical fields, and broadening access to quality jobsfor a diverse workforce. Table of Contents Introduction...............................................................................................................................................4Methodology............................................................................................................................................6Findings..........................................................................................................................................