Universities across the country are investing in artificial intelligence cultivation. Students participate in workshops, experiment with generative AI tools, and earn certificates intended to show they are ready for an economy shaped by AI.
The problem is that we don’t know whether these programs help students solve career problems, open doors to job opportunities, or both. We also don’t know if they give students from different institutions a fair chance to compete.
Higher education has an AI pilot program problem. A large part of the evidence of “successful” programs is still coming from single-region studies and relies on short-term measures such as test scores and self-reported data. Too often, they are described as successful before anyone can demonstrate whether students actually used what they learned outside of the classroom.
Colleges should stop viewing AI mastery as a finish line and start establishing clear pathways from learning to workplace application. Every AI introductory program should be designed with employers, tested across multiple institutions, and evaluated based on what students can actually do next. Attendance, certificates and satisfaction are not enough. The goal should be preparation.
Related: Interested in innovations in higher education? Subscribe to our free bi-weekly higher education newsletter.
Mastery of AI is necessary. Yet just knowing how to use a tool does not mean a student knows how to responsibly apply it to an unfamiliar workplace problem. A student may know how to ask a chatbot a question but not know how to verify the answer, protect private information, explain a recommendation, or decide when the tool should not be used. These abilities develop through repeated practice in real-world contexts. This is why colleges and universities should adopt common AI workforce readiness measures.
The stakes are particularly high for students at public and less resourced institutions. They may be introduced to AI tools without receiving enough practice, guidance, or employer connections to turn that introduction into an opportunity.
Earlier this year I founded an initiative to examine this challenge in two different contexts. The program brought together CUNY undergraduates and NYU Tandon engineering graduate students in the same series of hands-on workshops. Industry professionals showed students how AI is used in real jobs, and students practiced using the tools for workplace tasks and explaining their decisions.
Our little one, preliminary study found that both groups achieved similar levels of applied AI use. Participation emerged as the clearest factor in predicting this use. Among students who participated in one or two workshops, regardless of school, 21% created something with AI. Of those who participated six or more, 56 percent did so. The study merits replication elsewhere, but it did not detect a difference between the two institutional groups on this measure. Curriculum design, as well as the amount of meaningful practice students receive, may be of even greater importance.
Related: OPINION: Schools can’t teach AI without a way to measure it
A useful question is whether particular combinations of teaching, practice, career guidance, and real-world opportunities produce meaningful results, and which students benefit from them. To answer this question, universities will need to go beyond isolated pilot projects.
First, to evaluate their AI induction programs, colleges should measure changes in student judgment, problem solving, professional communication, and the ability to create meaningful work with AI. They should also check whether students have access to internships or jobs.
Second, universities should study programs at different types of institutions. We cannot automatically assume that the results of a given program generalize to different institutional contexts.
Third, employers should become partners in the design of pilot programs rather than occasional guest speakers. They can help define real problems, review student work, and explain what abilities are important on the job. Students need connections to professionals who can show them what responsible use of AI looks like in a real organization.
It is encouraging to note that federal labor policy East begins to move in this direction. In July, the U.S. Department of Labor receives nearly $162 million through five agreements to expand registered apprenticeship programs. The funding uses performance-based incentives linked to outcomes such as hiring new apprentices, retaining them in programs and progressing.
Jobs for the futurea national nonprofit, for example, received $40 million to support the growth of registered apprenticeships in roles building and maintaining critical infrastructure that supports the artificial intelligence, semiconductor and nuclear energy industries.
But apprenticeships are not college courses: They pay salaries, provide federally recognized credentials, and place the apprenticeship in the context of employment. While colleges can’t simply copy this model, they can pursue the same goal: training that leads to demonstrated skills, meaningful employer engagement, and measurable outcomes.
Since generative AI reached college campuses, universities have shown that students are interested in learning how to use it. The next phase should determine which programs actually help them use AI responsibly and integrate these skills into the workplace. Future grants should require shared metrics, cross-campus comparisons, and follow-up after students leave, because no single campus can establish this evidence.
Without this evidence, AI workforce readiness will remain a collection of promising campus stories. Through this, higher education can build a system that works for students, regardless of where they enroll.
Ngoc Cindy Pham is Associate Professor of Marketing at Brooklyn College, CUNY, and Visiting Research Professor at NYU Tandon. She is also a Fulbright specialist and founder of BRIDGE AI Lab.
Contact the opinion editor at opinion@hechingerreport.org.
This story about college and AI was produced by The Hechinger reportan independent, nonprofit news organization focused on inequality and innovation in education. Register with Hechinger weekly newsletter.
The post OPINION: Knowing how to use AI tools is not the same as knowing how to apply them responsibly. Colleges Need to Do More appeared first in the Hechinger Report.
