Are We Helping Workers Reskill for the Future of Work? Using AI to Explore the Alignment of Online Course Offerings...

Millions of workers and job seekers turn to online platforms to gain work-relevant skills to remain competitive for the future of work. However, little is known about whether the skills acquired in work-relevant online courses align with the skills required for 21st-century jobs. Drawing on literature on job and skill matching, this exploratory study examines the alignment between available online training and learning content and the skills demanded by jobs (i.e., training-skills demands fit) using artificial intelligence methods. A large language model (LLM; Claude Haiku 3.5) was instructed to evaluate which of the 35 basic and cross-functional skills from the Occupational Information Network (O*NET) could be acquired in a given course, which was based on 2549 course descriptions extracted from MIT OpenCourseWare. Linkages between online training and skills were broken down by job family and occupations with a bright outlook designation (i.e., occupations estimated to have 75,000 or more job openings between 2024 and 2034 across the United States). Results suggest that the skill of active learning (i.e., using new information for problem-solving; 88%, N = 2242) was linked to the highest number of online courses, whereas the skill of instructing (i.e., teaching others to perform tasks; 5.3%, N = 134) was linked to the least. Computer and mathematical occupations had the highest proportion of courses wherein individuals can acquire basic and cross-functional skills, whereas food preparation and serving occupations had the lowest proportion of courses. Non-bright outlook occupations had a significantly lower proportion of online courses where individuals can acquire basic and cross-functional skills compared to occupations with a bright outlook designation. We expand on existing skills-matching perspectives to consider how training-skills demands fit can constrain or facilitate continuous learning and development. Further, we illustrate how LLMs can be used to efficiently and at scale summarize descriptive information on talent development issues.