# Career pathways are growing but evidence isn’t > Source: > Published: 2026-08-19 12:00:00+00:00 Demand for [AI](https://www.fastcompany.com/section/artificial-intelligence) skills in entry-level jobs nearly [tripled](https://www.naceweb.org/job-market/trends-and-predictions/demand-for-ai-skills-in-entry-level-jobs-nearly-triples-since-fall-2025) since last fall. High school seniors are graduating into that market now. Schools and states are responding by building new career pathways at a pace nobody would have predicted five years ago. Whether any of it works remains an open question. Tracking outcomes in career-connected learning is one way to find out. Career-connected learning is a continuum that includes career awareness starting as early as middle school, exploration through job shadowing and internships, and preparation through career and technical education and work-based learning. However, a recent Bellwether and Britebound [report](https://www.britebound.org/insights/making-it-work-report) profiles work-based learning efforts in 10 states and finds that each state is taking a different approach with different definitions of what counts, different data collected, and different infrastructure for tracking it. Even in states collecting data, few can say which approaches move outcomes, for which students, and under what conditions. [Nebraska](https://www.alicoalition.org/blog/how-federal-rd-investments-are-strengthening-cte-programs-and-improving-student-outcomes-in-nebraska/) shows what’s possible when strong evidence infrastructure exists. A federally funded [evaluation](https://ies.ed.gov/sites/default/files/rel-central/document/2025/10/REL_2021087.pdf) through the Regional Educational Laboratory found that students in Nebraska and South Dakota who completed structured career-technical education (CTE) pathways were seven percentage points more likely to graduate on time and 10 percentage points more likely to enroll in postsecondary education within two years. That evidence gave Nebraska lawmakers what they needed to authorize the state’s first-ever dedicated CTE appropriation, and CTE is now offered in every public high school in the state. However, the same study shows data’s limits. The researchers set out to measure whether CTE students earned more income after high school but couldn’t because there is no infrastructure linking education records to workforce data. One of the best evaluation programs failed to answer the question that matters the most to students. Evidence can change outcomes when there is a working system to produce it, carry it, and act on it. For career-connected learning, that system barely exists at the scale needed today. The federal government has an essential role to play in building that system. Last month, our team at the Alliance for Learning Innovation (ALI) released a [federal policy brief](https://www.alicoalition.org/publications/ccl/) that lays out how. We also [briefed Congress](https://www.edweek.org/teaching-learning/students-want-career-education-more-research-can-improve-it-new-report-says/2026/06). State leaders, researchers, and employers walked policymakers through the current evidence—or lack of it—surrounding career-connected learning. The brief, shaped by field experts and ALI coalition members, calls on the federal government to act on these three recommendations. Most states still can’t follow their students from K-12 into postsecondary education and the workforce, which means they can’t tell which career-connected learning experiences are paying off. States like [Kentucky](https://kystats.ky.gov/Reports/Reports) and [Colorado](https://ed.cde.state.co.us/slds) have preschool through workforce longitudinal data systems that measure student outcomes and use what they learn to inform everything from policy to classroom instruction. The federal government is well positioned to do the same. The field must be able to study what works – and the federal government should take a leading role. That could mean creating an Advanced Research Projects Agency ([ARPA](https://emergingtechpolicy.org/institutions/executive-branch/arpas/))-inspired Education and Workforce R&D Center with the [hiring](https://www.fastcompany.com/section/hiring) flexibility, funding tools, and program-manager autonomy that made the ARPA model successful in defense and energy. It might also look like a Career-Connected Learning Panel at the [Institute of Education Sciences](https://www.alicoalition.org/ncade-faq/) (IES), built directly on the [National Reading Panel](https://www.nichd.nih.gov/research/supported/nrp) model to evaluate the causal evidence we have and tell states and districts which approaches to scale. The National Science Foundation and the Departments of Labor (DOL), Commerce, and Defense also touch career-connected learning, but their investments aren’t widely coordinated. For example, the AI Workforce Research Hub, led by the DOL, Commerce, and the Bureau of Labor Statistics, is designed to evaluate how AI and automation are reshaping the American workforce and translate those findings into insights that states and schools can use. Efforts like these need to become the norm, with the federal funding to match. U.S. Rep. Suzanne Bonamici (D-Ore.) made it clear during ALI’s Annual Congressional Briefing this past June that students and teachers are navigating a rapidly changing world without a map. She argued that the federal government has helped provide that map for other sectors through sustained investment in R&D for defense, health care and energy. It’s time they make the same commitment to education. She’s right. States aren’t waiting for permission. They’re building pathways with the best information they have, which isn’t much. One of the strongest studies in the field went looking for the most important answer and found the data had never been collected. We can give them better. *Sara Schapiro is the executive director of Alliance for Learning Innovation.*