In 2017, Jason Swanson laid out several possible futures for work and argued that schools should focus on self-knowledge and learning skills. Swanson’s article appears below as published. After it, a new section checks Swanson’s questions against current United States labor data and asks what schools can sensibly do.
How do we prepare students for jobs that don’t even exist yet?
To add potentially more stress to it, the future of work feels increasingly complex and uncertain. There is quite a bit of speculation as to what impact factors such as artificial intelligence (AI), automation, globalization, and taskification (the breaking up of large jobs into small tasks) might have on the future of work. For example:
- Might we experience an abundance economy where artificial intelligence and automation drive consumer prices down and eliminate boring and dangerous work?
- Might most people be forced into contingency employment, competing with one another to assemble a mosaic of paid work?
- Might the vast majority of people find themselves out of work as the march of automation accelerates and human workers are replaced faster than new jobs can be created?
- Might work be redefined as artificial intelligence and automation eliminate the majority of jobs, leading people to work on task and projects based on their passions rather than on the need to generate income?
In reality, the future will probably contain shades of all the above, along with other developments that we have not yet considered.
Trends Shaping the Future of Work
Even if we take drivers of change such AI and automation out of the mix, the world of work is still in flux. The average worker in the US now holds 11.7 jobs in his or her lifetime, an excellent reminder that most of us already experience a fair degree of churn. We can expect that number to rise as the way people organize work changes. One factor that will likely contribute to such an increase is the rise of the contingency-based workforce. McKinsey Global Institute estimates that 54 to 68 million people in the United States already work in the project-based economy (i.e., as contingency-based employees). This number is expected to increase thanks to factors such as the lower coordination costs provided by the Internet and the continued proliferation of digital platforms that help match candidates with organizations seeking to complete a task.
Editor’s note (2026): The current Bureau of Labor Statistics release gives 12.9 jobs held between ages 18 and 58, which is a different group of workers from the 11.7 figure cited here.
Best if Used by…
These expected changes point toward a future of work where the shelf life of skills will be shortened dramatically. The changing structure of jobs might mean that one indiviudal could be doing very different things from one project to the next. Bigger picture, the tasks in which we engage and the nature of our contributions to work will likely change quickly as machine partners do more and more for and with us. We will be likely to need just-in-time training to keep up with work’s changing demands, with new skills enduring for far less time than they do today.
A Hardware Upgrade
One answer might be by helping them cultivate deep self-knowledge and meta-cognitive skills. In other words, focus on upgrading a learner’s hardware (how to think, learn, and unlearn) rather than their software (skill acquisition and knowledge recall), promises to be the best way of ensuring that learners are ready for a future of work, no matter which future of work comes to fruition.
Due to the uncertain nature of future work, skills and practices such as deep self-knowledge, social awareness, emotional regulation, self-determination and proactive learning will become increasingly important. For instance, in a future of work where employment structures are increasingly contingent, a learner will need such skills to manage time if they are working on many different projects, to seek skills to complete those projects and to have the emotional intelligence to partner with both people and machines from across the globe. In a post-work world, such skills might be equally important, helping people create subsistence strategies, pursue community-impact initiatives, and start passion-based projects rooted in their values.
Cultivating core social-emotional skills and foundational cognitive and meta-cognitive practices also promises to assist individuals in creating bold aspirational visions for themselves and in navigating terrain that today can only look uncertain, even uncharted. Focusing on these skills can provide the foundation for both near-term career pathways and future career creation and success.
Preparing students for future jobs in 2026: what labor data shows and what schools can do
Preparing students for future jobs is harder when nobody agrees on what those jobs will be. As Jason Swanson explains above, artificial intelligence, automation and contingent work could lead in several directions, and Swanson’s advice was to build the skills that apply in any of them. Nine years on, government data offer a clearer view of some of those questions. They do not settle the debate, but they give parents, teachers and school leaders firmer ground than speculation.
What the current projections say
The US Bureau of Labor Statistics published its employment projections for 2025 to 2035 in August 2026. It projects 5.9 million new jobs, with total employment rising from 170.3 million to 176.2 million, growth of 3.5 percent. That is slower than the 10.9 percent growth recorded over 2015 to 2025. The fastest-growing occupations include nurse practitioners at 41.0 percent, solar photovoltaic installers at 36.5 percent and data scientists at 34.6 percent.
The release also addresses the question Swanson raised about automation. BLS says AI adoption is expected to support some occupations, such as those tied to data science and AI development, while productivity gains from AI and automation may reduce demand for others. Office and administrative support is projected to decline fastest among major occupational groups, by 4.0 percent, losing 752,100 jobs. BLS says generative AI may limit demand in arts, design, entertainment, sports and media occupations. The pattern sits between the two views Swanson sets side by side, in which automation either redefines jobs, as James Bessen argues, or removes them faster than new ones appear. The projections show a shift within particular occupations over ten years, not the wholesale loss that some of Swanson’s scenarios imagined.
Parents and teachers can use this when students ask which jobs are safe. A fair answer is that no list is guaranteed, that health care occupations and solar and wind energy jobs are among those expected to grow, and that the same tools that reduce demand in some office roles are creating work in others. Students who can learn a new tool, check its output and explain their reasoning are in a stronger position in either case.
How often people change jobs
Swanson cited a figure of 11.7 jobs in a lifetime. The BLS National Longitudinal Survey, which follows the same people over time, reported in 2025 that individuals born from 1957 to 1964 held an average of 12.9 jobs from ages 18 to 58, including 5.6 jobs between ages 18 and 24. These are different cohorts and measures, so the numbers should not be compared as a trend. They do support Swanson’s central point that most people change jobs many times, and that much of the changing happens early in a working life, when students are still deciding what to learn.
What schools can reasonably do
Swanson’s recommendation was to upgrade the learner’s “hardware,” meaning how to think, learn and unlearn, instead of focusing only on skill acquisition and knowledge recall. The labor projections do not contradict that, but they also show that specific knowledge matters in growing fields such as health care, energy and data. The writer’s reading, which is a judgment and not a finding of any study, is that schools need both: strong core knowledge and the habits that let a graduate keep learning.
In practice, a few decisions follow. Teachers can give students work where they plan a long task, check their own progress and revise, since those routines build the self-management Swanson describes. Schools can offer students repeated, structured contact with people who work in different fields, so that choices rest on more than a guess. Students can practice working with AI tools as an assistant, and learn to check the output, rather than being kept away from them. Counselors and teachers can show how a first job often leads to others, using the survey figures above.
Leaders should also be cautious about claims. Predictions that a given job will vanish, or that a certain skill will be essential, have often proved wrong, and the BLS itself describes AI effects as expectations that may change. A school that teaches students to read such projections critically is teaching a skill that will last longer than any prediction.
Related edCircuit guides
Readers planning how schools approach artificial intelligence can turn to the AI in education guide and the grade-by-grade guide to AI literacy for students.
Further Reading
- USA Today – White House: Workforce development aimed at apprenticeships, future jobs
- VentureBeat – Udacity: Reno pilot shows the future of tech job training
- The Detroit News – Preparing for the jobs of the future
Related reading
More on Career and Technical Education (CTE)
