Two years into the generative AI boom, the mass layoffs many workers feared haven't appeared in the numbers.
That's the central finding echoed by labor economists who say the gap between AI anxiety and AI's measurable impact on employment remains wide.
Unemployment directly attributable to automation is still difficult to isolate in national employment data, even as executives mention AI in earnings calls with increasing frequency.
The disconnect matters because it shapes how workers, employers and policymakers respond.
Overreacting to a threat that has not fully materialized could lead workers to abandon viable career paths or companies to automate roles that still require significant human judgment.
But the opposite response carries risks too.
The question may be less whether AI takes jobs than how quickly it changes the jobs people already have.
Slow Burn, Not a Cliff
Economists increasingly describe AI's labor impact as gradual rather than sudden.
Jobs are being restructured task by task rather than eliminated wholesale.
A customer service representative might handle fewer routine calls while taking on more complex cases.
A paralegal might spend less time reviewing documents and more time supporting strategy.
That's a meaningfully different story from the one dominating some headlines and corporate presentations.
Historical parallels offer some perspective.
The introduction of ATMs did not eliminate bank tellers. It changed what they did and ultimately contributed to changes in how banks staffed branches.
Spreadsheet software did not wipe out accountants. It shifted the value of their work away from arithmetic and toward analysis.
AI could follow a similar pattern, although its ability to perform cognitive tasks makes the comparison imperfect.
Where the Real Risk Lives
None of this means AI's impact on employment will be negligible.
Entry-level roles in areas such as coding, writing and basic data analysis may be particularly exposed because AI systems can increasingly perform routine versions of those tasks.
Younger workers and new graduates could feel that pressure first.
That would represent an important difference from some previous technological transitions, which often affected manual or routine physical work before highly educated white-collar roles.
There is another potential problem: wages.
Even if jobs survive, AI could increase worker productivity without necessarily increasing worker bargaining power.
That could put downward pressure on wage growth in some occupations, particularly where employers can substitute AI-assisted workers for larger teams.
There is also a data-lag problem.
Labor statistics take time to capture structural changes, meaning today's employment numbers may not fully reflect disruption already occurring inside individual companies.
Why the Debate Is Changing
The timing of this reassessment matters.
Corporate AI adoption has moved beyond experimentation into deployment across sectors including finance, healthcare, legal services and customer support.
Yet productivity growth has not surged in proportion to the scale of AI investment that many expected.
That gap between AI spending and measurable economic output has become part of a broader debate about the productivity impact of the technology.
If AI were already eliminating large numbers of jobs, economists would expect at least some of that disruption to become visible in employment and productivity data.
So far, the evidence does not show a straightforward collapse in employment attributable to AI.
That could change.
It simply hasn't happened at the scale many early predictions suggested.
What Changes From Here
The debate is likely to shift from "Will AI take my job?" to "How will AI change my job?"
That reframing matters for workers deciding what skills to develop, companies restructuring teams and policymakers designing workforce programs.
Companies betting heavily on immediate AI-driven headcount reductions may discover that the financial payoff is slower and more complicated than promised.
Businesses using AI as a productivity multiplier for existing employees may have a more practical path to near-term value.
The distinction is important because automation and augmentation are not the same economic strategy.
One replaces tasks.
The other changes who performs them, how quickly they are performed and what workers spend their time doing.
What This Means for Miami
South Florida's economy is heavily exposed to sectors including hospitality, real estate, healthcare, finance and logistics.
Many of those industries contain roles where judgment, relationships, physical presence and local knowledge remain important.
That does not make them immune to AI.
It means the more immediate opportunity may be augmentation rather than wholesale replacement.
For Miami's growing technology and startup community, that distinction is strategically important.
Companies building AI products may find stronger enterprise demand by helping workers perform existing jobs more efficiently rather than promising to eliminate entire departments.
Local employers experimenting with AI can likewise test specific use cases without assuming that adoption automatically requires immediate headcount reductions.
For Miami's universities and workforce-development programs, the implication is equally straightforward: preparing people to work effectively with AI may matter more in the near term than preparing them for a world in which AI simply replaces them.
The labor market may not be facing an AI cliff.
It may be facing something more gradual — and potentially harder to see coming: a slow restructuring of what employers expect people to do.