This paper is the first of two. The second paper is available here. These papers will form the basis of our formal response to the Department of Labor’s Request for Information. We’re publishing them now because we believe the ideas improve through discussion, and we welcome thoughtful feedback before submitting our final comments.
Response to the Request for Information on Modernizing the Occupational Information Network (O*NET)
To be submitted to the U.S. Department of Labor, Employment and Training Administration (once feedback is incorporated)
Please provide your feedback using this form.
Modernizing Workforce Infrastructure for the AI Era
The greatest challenge facing workforce systems is no longer understanding work. It is understanding people.
The Occupational Information Network (O*NET) has served the United States well as the nation’s authoritative source for describing occupations. For more than two decades, it has provided the common language used by workforce agencies, educators, employers, researchers, and technology providers to understand work.
We strongly support the Department’s effort to modernize O*NET. The questions posed throughout this Request for Information recognize an important reality: occupations are evolving more rapidly than ever before, driven by artificial intelligence, technological innovation, demographic change, and shifting labor market demands.
We believe, however, that the opportunity before the Department is even larger than modernizing occupational data.
Artificial intelligence has fundamentally changed the economics of workforce information. Systems are no longer limited by their ability to store occupational descriptions. Increasingly, they are limited by their ability to understand and reason about human capability.
The nation’s ability to respond to technological disruption increasingly depends on whether workforce systems can recognize what people are capable of—not simply where they have worked or what degrees they have earned.
Historically, workforce systems have been organized around occupations. Individuals were expected to navigate toward occupations, education aligned itself to occupations, and workforce programs measured success through occupational outcomes. That model reflected the limitations of previous technology.
Today’s intelligent systems operate differently. They reason across capabilities, experiences, credentials, tasks, and evidence. They compare people to opportunities, not simply occupations to occupations. Yet our workforce systems continue to understand people primarily through resumes—documents originally designed to summarize past employment rather than represent human capability. As AI makes it easier to optimize resumes around job descriptions, those documents increasingly reflect what applicants believe employers want to see instead of providing an accurate representation of what people actually know and can do.
For this reason, we believe the next generation of workforce infrastructure should separate the representation of human capability from the representation of occupations.
O*NET should remain the nation’s authoritative representation of work. But it should no longer be responsible for defining the foundational language of human capability upon which occupations are built.
I. The Workforce System Must Learn to Understand People Differently
For much of the twentieth century, the workforce system was built around a relatively stable model of education and employment. Individuals acquired most of their formal education early in life, entered an occupation or industry, and developed through a sequence of related jobs. Degrees, job titles, and years of experience were incomplete representations of a person, but within a more stable labor market they often provided enough information for institutions and employers to make decisions.
That paradigm no longer holds.
Workers are changing jobs and careers more frequently, while technology is changing the tasks performed within occupations themselves. People increasingly need to learn throughout their lives, develop new capabilities, and move between roles that may not share the same title, industry, or traditional qualifications. Workforce systems must therefore do more than help people prepare for an initial occupation. They must help individuals understand what they can already do, what capabilities remain valuable, what they need to learn next, and where those capabilities can lead.
Yet the systems responsible for supporting these transitions still understand people primarily through two proxies: educational qualifications and work history.
Both are becoming less sufficient.
A degree communicates where and what someone studied, but often reveals little about the specific capabilities they developed, retained, or subsequently added. A job history describes where someone worked and the roles they held, but not necessarily the tasks they performed, the problems they solved, the proficiency they developed, or the capabilities they can transfer into a different context.
Artificial intelligence makes this limitation more urgent. As AI automates, augments, and reorganizes work, prior experience may become less predictive of future opportunity. A person may have spent years performing tasks that are now automated while also possessing judgment, knowledge, relationships, and transferable capabilities that remain highly valuable. Job titles and resumes alone cannot reliably distinguish between the two.
At the same time, AI is increasing the scale at which people may need to retrain and transition. If workforce systems cannot recognize the capabilities individuals already possess, each transition risks forcing people to begin again: repeating learning, recreating evidence, and being evaluated only against the requirements of the next job rather than the full value they have accumulated throughout their lives.
That outcome would be costly for workers, employers, educational institutions, and government. It would also waste enormous amounts of human capability at the moment the economy most needs to redeploy it.
The modernization of workforce infrastructure must therefore begin with a more sophisticated way of understanding people. The nation needs a shared language capable of describing human capability independently of the degree through which it was taught, the job in which it was demonstrated, or the institution that recorded it.
That shared language is the foundation upon which a National Human Capability Ontology should be built.
II. Establish a National Human Capability Ontology
Modern workforce infrastructure requires a shared public language for describing human capability.
Today, concepts such as skills, knowledge, abilities, competencies, tasks, tools, behaviors, and credentials are embedded throughout occupational profiles and duplicated across education systems, credential registries, employer technologies, and workforce platforms. While these systems often describe the same underlying capabilities, they frequently do so using different structures, identifiers, and terminology.
The result is unnecessary fragmentation. Every organization spends considerable effort translating between representations rather than improving how individuals discover opportunity.
We recommend that the Department establish a National Human Capability Ontology as foundational public infrastructure.
The ontology would define the foundational concepts and relationships that describe human capability, including skills, knowledge, abilities, competencies, tasks, behaviors, tools, experiences, credentials, and evidence. Rather than embedding these concepts separately within every workforce application, they would exist as reusable public infrastructure that other systems reference.
III. Reposition O*NET as the Nation’s Representation of Work
The creation of a National Human Capability Ontology does not replace ONET any more than a dictionary replaces a book. The ontology provides the shared vocabulary. ONET provides the authoritative description of how those capabilities combine within occupations. Both are essential, but they serve fundamentally different purposes.
The National Human Capability Ontology defines the language. O*NET tells the story of work using that language. Establishing a National Human Capability Ontology does not diminish the importance of O*NET. It clarifies its role.
O*NET should remain the nation’s authoritative representation of occupations. The difference is architectural. Today, occupational profiles largely define the capabilities associated with each occupation. In the future, occupational profiles should instead reference capabilities defined within the National Human Capability Ontology.
Occupations become compositions of reusable capabilities rather than containers that redefine them. This seemingly modest architectural change has significant implications. It eliminates duplication across occupational profiles. It enables capabilities to evolve independently of occupational classifications. It allows emerging occupations to be assembled from existing capability definitions while introducing new concepts only when necessary. It enables AI systems to reason consistently across occupations because every occupation references the same underlying language.
Most importantly, it separates two fundamentally different responsibilities. The ontology becomes responsible for representing human capability. O*NET becomes responsible for representing how those capabilities combine to perform work. This separation creates a more flexible architecture while preserving O*NET’s essential role as the nation’s trusted occupational reference.
IV. A Foundation for Workforce Innovation
The value of this approach extends well beyond O*NET itself. A National Human Capability Ontology would provide a shared language that could be adopted across federal agencies, states, educational institutions, employers, standards organizations, and technology providers.
Credential providers could describe learning outcomes using the same capability definitions referenced by occupations. Education providers could map curriculum directly to nationally recognized capabilities rather than creating institution-specific interpretations. Employers could describe work using language consistent with workforce programs and educational pathways. Researchers could compare labor market trends using a common semantic model. Artificial intelligence systems could reason across people, occupations, credentials, and opportunities using a shared public vocabulary rather than proprietary mappings.
Most importantly, future workforce systems would no longer need to choose between describing people or describing work. Both would reference the same foundational language. O*NET would describe occupations. Other systems would describe individuals. Because both rely upon the same public infrastructure, interoperability becomes a natural consequence of the architecture rather than an ongoing integration challenge.
Conclusion
The Department has an opportunity to do more than modernize O*NET. It has the opportunity to modernize the architecture upon which workforce information is built. For more than twenty years, O*NET has provided the nation with a trusted representation of work. The next generation of workforce infrastructure requires an equally trusted representation of human capability.
We therefore recommend that the Department establish a National Human Capability Ontology as foundational public infrastructure and position O*NET as the nation’s authoritative occupational representation built upon that foundation.
Such an architecture preserves the strengths that have made O*NET successful while creating the flexibility, interoperability, and AI-readiness required for the decades ahead.
Just as ONET became the nation’s trusted representation of work, the next generation of workforce infrastructure should give the nation an equally trusted representation of human capability.
Signol Labs 
Leave a comment