For more than a generation, Americans were given remarkably consistent career advice: Learn technology. Get a computer science degree. Become a software engineer. Learn to code. While manufacturing jobs disappeared and entire industries moved overseas, technology was presented as the safe harbor—the sector where education, intelligence, and specialized skills would provide protection against economic disruption.
Artificial intelligence may be about to test that promise.
The uncomfortable possibility is that some of the people most vulnerable to AI are precisely those who were told they were preparing themselves for the future. Programmers, software developers, data analysts, cybersecurity specialists, graphic designers, engineers, and other highly educated technical professionals increasingly confront machines capable of performing portions of their jobs at extraordinary speed.
This does not mean millions of technology workers suddenly become unemployed. Technological revolutions rarely happen that cleanly. But AI does not need to eliminate a profession to dramatically disrupt it. It merely needs to allow one employee to accomplish what previously required three, five, or ten.
Consider software development.
A programmer once had to write thousands of lines of code manually, search documentation, debug errors, construct tests, and spend hours locating an obscure problem. Increasingly capable AI systems can already generate code, explain unfamiliar codebases, identify bugs, write tests, create documentation, and translate instructions into functioning software.
The experienced engineer does not necessarily disappear. Instead, that engineer becomes enormously more productive.
And therein lies the problem.
If ten developers assisted by AI can accomplish what once required 30 developers, what happens to the other 20?
For years, technology optimists have responded to automation concerns with a familiar argument: Technology eliminates some jobs but creates others. Historically, they have a powerful case. Automobiles destroyed occupations associated with horses while creating enormous industries surrounding manufacturing, highways, petroleum, trucking, repair, insurance, and tourism. Computers eliminated countless clerical tasks while creating industries that previous generations could scarcely imagine.
AI, however, presents an unusual complication. The machine is increasingly capable of learning portions of the new jobs too.
When manufacturing automation displaced physical labor, workers were encouraged to move toward knowledge work. When routine office work became automated, Americans were encouraged to acquire technical skills. But where does a software engineer go when software engineering itself becomes substantially automated?
That question deserves more serious consideration than another government-funded retraining program.
The answer may begin with recognizing that a college degree is not economic armor. America has spent decades encouraging young people to acquire credentials while simultaneously treating skilled trades as consolation prizes. AI could expose how misguided that hierarchy became.
The electrician repairing a century-old house, plumber diagnosing a broken pipe, HVAC technician replacing a compressor, mechanic rebuilding machinery, and lineman restoring electricity after a storm possess something increasingly valuable: They operate in an unpredictable physical world.
AI can tell someone how to replace a water heater. It cannot easily crawl into a cramped basement and replace it.
That does not mean every displaced programmer should become an electrician. It means America should stop assuming that occupations requiring four-year degrees inherently possess greater economic security than occupations requiring physical expertise.
Technical professionals themselves will also adapt. The strongest programmers may become architects and supervisors of AI-generated software. Cybersecurity specialists may increasingly defend systems against AI-generated attacks. Engineers may oversee automated design. Entrepreneurs may use AI to build companies with staffing levels previously impossible.
But adaptation will create winners and losers.
The greatest danger would be Washington attempting to freeze the labor market in place. Politicians will inevitably face pressure to protect threatened professions through subsidies, regulations, licensing requirements, artificial staffing mandates, or restrictions on automation. Such policies may preserve particular jobs temporarily, but they would also make American businesses less competitive against countries that embrace the technology.
America cannot regulate itself back into 2015.
Nor should the answer automatically be universal basic income or some enormous permanent government program paying displaced professionals not to work. Human beings need more than checks. Work provides independence, purpose, achievement, social connection, and dignity. A society that responds to technological abundance by placing millions of capable people on permanent government dependency would solve an economic problem by creating a cultural one.
The better response is mobility.
Government should remove barriers that make changing careers unnecessarily difficult. Occupational licensing should be reconsidered where it serves primarily to restrict competition. Apprenticeships and technical education should become genuine alternatives to universities. Colleges should be forced to demonstrate that expensive degrees actually produce economic value. Tax and regulatory policy should make entrepreneurship easier, not harder.
Most importantly, Americans may need to abandon the notion that education ends at 22.
The AI economy could require repeated reinvention throughout adulthood. A programmer may become an AI systems manager, entrepreneur, technical consultant, robotics specialist, instructor, or something that does not yet have a name.
There is considerable irony here. The workers who built the digital economy may become among the first professionals forced to confront its next great disruption.
That should not inspire panic. But neither should Americans accept Silicon Valley’s soothing assurances that entirely new categories of employment will magically appear in precisely the numbers required.
AI could create enormous prosperity while simultaneously destroying familiar career ladders. Both things can be true.
The defining economic question of the AI revolution may therefore not be whether machines can perform our jobs.
It will be whether a free society can create enough opportunity for human beings to keep finding better ones.