The federal government should advocate for AI-based skills assessments tailored to high-demand jobs.
For many years, Americans have viewed a college degree as a reliable indicator of competence. This perception persists not due to its accuracy, but rather a lack of superior, scalable alternatives. Artificial intelligence (AI) holds the promise of dismantling the so-called “paper ceiling” that degree requirements create, enabling businesses to evaluate worker capabilities swiftly and accurately—regardless of whether those skills were acquired through formal education or life experiences.
The urgency for AI-based skills assessments is underscored as the labor market evolves, becoming increasingly fluid and project-oriented due to technological advancements. A growing number of Americans are engaging in side jobs and pursuing freelance opportunities to gain better control over their schedules, advance their careers, and utilize new AI technologies. Employers, on their part, are shifting toward hiring for targeted, short-term projects instead of onboarding permanent employees, who may leave within a year or less.
In this changing environment, employers express frustration over the difficulty in finding qualified candidates, while prospective workers find themselves excluded from opportunities for which they are qualified. The reliance on credentials, which are becoming less dependable, incurs costs for everyone involved in the labor market.
AI has the potential to address this information gap—the uncertainty surrounding whether a degree genuinely signifies competence and whether experience alone can establish job readiness.
The key factor in effective labor matching is skill mastery—not merely credit hours. This principle has always held true, although it is more evident in certain sectors. In fields such as software, aviation, logistics, and skilled trades, employers frequently prioritize demonstrated competence over degrees.
Canada’s Red Seal Program exemplifies how well-crafted, transferable assessments can validate skills across various provinces and employers. Under this program, tradespeople only need to pass a standardized exam to show they are adequately trained in their field. The program encompasses roles in construction, automotive and mechanical work, manufacturing, landscaping, and the service sector. In these areas, the job market operates more efficiently because skills can be observed and evaluated directly.
In contrast, the United States has not genuinely attempted to replicate this logic across high-demand occupations. Instead, reliance on degrees—often expensive, time-consuming, and poorly aligned with job roles—has persisted as the primary signal of qualifications available. This approach inflates hiring criteria, marginalizes capable candidates, hinders mid-career transitions, and slows down economic adaptability.
This is where AI could be beneficial: by developing trustworthy, job-relevant assessments that employers value, while maintaining choice, encouraging merit-based competition, and reducing the disproportionate emphasis on educational credentials that are increasingly ineffective.
The question remains: should these tools be created and implemented? In theory, companies should be eager for tools that streamline applicant evaluations, as they can simplify a generally lengthy and costly process, aligning their needs with workers’ skills. Yet, AI adoption is lagging among private-sector firms, particularly smaller businesses that would benefit most from reducing hiring costs for top talent. This situation presents an opportunity for the federal government to act as a catalyst for applicable AI research and to adopt successful AI tools early in the process.
One potential approach could involve the U.S. Department of Labor initiating a challenge—not a mandate, a national exam, or a new bureaucracy—inviting industry players and academic researchers to design skills-based assessments for a select group of high-demand positions. Examples include industrial maintenance technicians, cybersecurity analysts, logistics coordinators, and data analysts—roles with concrete tasks and real shortages, where degree requirements often serve more as a filter than as a genuine assessment of skill.
The government would not dictate which assessment emerges as the “winner.” It would also not mandate employers to adopt any specific tools. Instead, it would undertake something more impactful: identifying safe, secure, and effective assessments, enabling better signals to be adopted widely.
An effective challenge would adhere to several core principles. First, job relevance: assessments must evaluate actual tasks—like troubleshooting systems, analyzing datasets, or responding to simulated incidents—not abstract trivia that AI can easily manage.
Second, there should be employer validation: success ought to be gauged not by flashy presentations but by whether employers in relevant fields agree that the assessments accurately predict job performance and are willing to use them in hiring or promotions.
Third, portability and openness are vital: workers should be able to transfer results across various employers, and no individual vendor should dominate the market.
Finally, accountability is crucial: if an assessment does not correlate with retention, productivity, or advancement, it should be deemed a failure—regardless of its technological sophistication.
AI makes this feasible in ways previously unachievable. It can simulate work environments, evaluate performance on a large scale, adjust assignment difficulty as applicants progress, and reduce assessment costs. Importantly, AI-based assessments would not replace human judgment; human resources professionals could continue to apply traditional screening methods if desired. These AI evaluations would merely serve as an additional source of information regarding an applicant’s capabilities.
Critics may justifiably fear potential overreach. A federal skills examination could pave the way for troubling scenarios. However, that’s not the intention here. The government already plays a role in uniting markets when coordination failures hinder progress, such as in setting technical standards or early internet protocols. Its role in this context would be similar—evaluating the effectiveness of tools and making them generally available while refraining from declaring winners.
Others might ask: why not leave this issue solely to employers? While theoretically feasible, no individual company is inclined to shoulder the entire cost of crafting a signal that others can benefit from without contributing. This is why degrees remain prevalent despite their shortcomings. A time-limited, competitive federal challenge could help address this collective-action problem without prescribing specific outcomes.
If executed correctly, this approach could lower entry barriers, expedite hiring processes, and make mid-career transitions more achievable. It could assist veterans in translating their experiences to civilian roles. It would provide workers with a means to demonstrate their current capabilities, rather than relying on outdated credentials. Ultimately, it could revitalize a credentialing system that has become rigid by default.
If these assessments fail—if employers disregard them, they don’t predict job performance, or they do not enhance hiring or career mobility—the initiative should be terminated. Even this outcome would yield valuable insights, reinforcing that degrees remain the least flawed signal in some areas. However, if even a few assessments succeed, the benefits could be considerable: quicker hiring, reduced entry barriers, more achievable mid-career shifts, and decreased dependency on imprecise educational requirements that often impede opportunities.
We have engaged in extensive discussions on whether a college degree is “worth it,” whether employers are overly demanding, and whether workers require more training. These conversations overlook a simpler bottleneck: the U.S. labor market is stagnant partly because we struggle to recognize skill acquisition and learning outside conventional pathways.
The Labor Department cannot address this challenge independently. However, it can facilitate solutions—by convening stakeholders, testing methodologies, measuring outcomes, and then allowing them space to develop. At a time when technological innovation is accelerating and career paths are increasingly nonlinear, establishing better methods to recognize skills is not a radical shift in policy; it is a necessary evolution in how we connect individuals with work opportunities.
