Remember the AI job apocalypse? The one where millions of white-collar workers would be replaced by algorithms, and unemployment would spike to Great Depression levels? Well, Bank of America has been digging through the data, and so far, the robots haven't exactly taken over.
Companies like Amazon (AMZN), Meta Platforms (META), Microsoft (MSFT), and Block (XYZ) have all cited AI-driven efficiencies and restructuring as they cut thousands of jobs. But when BofA economists looked at the broader labor market, they found little evidence that AI is causing a so-called "job apocalypse."
Instead, AI might be reducing demand for certain tasks and putting pressure on some white-collar and entry-level jobs, while simultaneously creating new employment in construction, manufacturing, and other sectors tied to the massive AI infrastructure buildout. It's a more nuanced picture than the doom-and-gloom headlines suggest.
The AI Job Apocalypse Has a Data Problem
BofA economists led by Stephen Juneau examined employment across 206 industries and compared job growth with each industry's exposure to AI. The results are striking. Industries with the highest AI exposure have seen employment largely move sideways since ChatGPT launched in late 2022. Less-exposed industries grew about 2%, and there is virtually no correlation between AI exposure and employment growth.
In simple terms, AI exposure explains almost none of the differences in job growth across industries. That weakens the idea that AI is already driving a broad employment collapse. BofA also found little relationship between AI usage and labor demand. Industries using AI more heavily have not consistently experienced larger declines in employment and job openings.
"AI replaces tasks not occupations," Juneau said. A worker may use AI to perform some tasks faster without making the entire job disappear. It's a subtle but crucial distinction.
Yes, But...
The picture is less reassuring in some white-collar industries. Information and finance and insurance, both with relatively high AI usage, saw labor demand decline in the five months through June, suggesting companies may already be turning to AI to reduce labor costs, BofA said.
Youth unemployment has also risen, particularly among recent college graduates aged 22–27. BofA said AI "may be contributing" to the relatively weaker labor outcomes for this group. That could be the first place where AI disruption has already become visible. In November, Stanford research found that employment among workers ages 22 to 25 in the occupations most exposed to AI has fallen 13% since 2022, suggesting AI may be contributing to weaker employment outcomes for young workers.
Still, AI investment is simultaneously creating jobs elsewhere. Nonresidential construction added 95,000 jobs year-to-date, while AI-related manufacturing added another 32,000. Together, BofA estimates those sectors account for roughly 25% of new private-sector jobs this year.
The AI infrastructure boom offers a real-world example of this trade-off. Vertiv Holdings Co. (VRT), which supplies power and cooling systems for data centers, is expanding its Ohio manufacturing footprint to meet demand for AI infrastructure. The company expects the project to create hundreds of jobs through 2029.
"AI capex is fueling job creation in the goods producing sectors," Juneau said. This is the part of the AI economy that the apocalypse narrative misses.
The Investment Question Is Changing
The real risk may not be that AI suddenly eliminates millions of jobs. It may be that AI changes where jobs are created. White-collar entry-level roles could face pressure while construction, manufacturing, power, and data-center infrastructure benefit from the enormous capital spending required to build AI capacity.
BofA's conclusion is therefore less "AI cannot destroy jobs" than "the destruction is not happening at the scale the market feared." For investors, that distinction matters.
Nightmare Scenario Remains Possible: Citrini
In February 2026, a 4,000-word memo from Citrini Research, titled "2028 Global Intelligence Crisis," imagined U.S. unemployment surging to 10.2% by June 2028 as AI displaced white-collar workers and triggered a damaging economic feedback loop. Citrini explicitly framed the memo as a scenario, not a prediction. But the thought experiment still gave investors a stark glimpse of what an AI-driven labor shock could look like.
Six months later, Bank of America is looking at labor-market data and determined that, for now, the AI job apocalypse remains a scenario waiting for evidence.
