The AI jobs debate has been apocalyptic, for the most part.
Anthropic CEO Dario Amodei warned AI might wipe out half of entry-level white-collar jobs, pushing unemployment toward 10% to 20%, Fortune reported. Tesla CEO Elon Musk went even further at the U.S.-Saudi Investment Forum, saying that “work would become optional,” Yahoo Finance noted.
Nvidia CEO Jensen Huang, though, took the other side, arguing that even though “many tasks will be automated away,” new jobs will emerge.
Even OpenAI CEO Sam Altman, who feared sweeping white-collar displacement, recently said that the feared global “jobs apocalypse” hasn’t quite materialized as swiftly as expected, according to Yahoo Finance.
That said, investors have been bracing for AI to replace workers. Yet in a note shared with me, Bank of America went looking for that destruction in the actual labor data.
The bank just released a new analysis spanning 206 industries, AI usage, job openings, younger workers, and the data-center investment boom. What it found could potentially reshape the debate over what AI is really doing to American jobs.
Is AI actually destroying American jobs?
Bank of America believes that despite all the gloom-and-doom scenarios presented by the Darios and Altmans of the world, the idea of AI triggering a white-collar jobs apocalypse doesn’t actually hold up.
BofA analysts argue that like any other technological disruption, AI is seemingly replacing tasks instead of entire occupations.
Workers can initially be displaced as businesses automate certain functions, but eventually new tasks and roles emerge to offset that damage.
More Layoffs:
- Samsung cuts jobs as it shifts U.S. headquarters
- Another popular soda giant closes warehouse operation, cuts 184 jobs
- Meta layoffs take disturbing turn in new lawsuit
Interestingly, the bank’s arguments are already backed up by the data.
BofA looked at jobs trends across 206 industries to see if sectors that were exposed to AI were losing jobs faster.
Those with the highest AI exposure have seen little to no job growth since ChatGPT hit the screens in late 2022. In comparison, those with the lowest AI exposure grew employment by nearly 2%.
However, the more telling thing emerges when we look at all 206 industries. BofA found no correlation between a sector’s AI exposure and its employment growth.
A similar conclusion is reached when looking at actual AI adoption.
Information technology businesses reported a 42.1% AI usage rate in June, and finance and insurance reported 34.8%. Across the broader economy, though, the bank saw little relationship between AI use and the combination of employment and job openings.
If hiring is getting worse, why has unemployment fallen?
Interestingly, the U.S. labor market has undeniably weakened.
Nonfarm payrolls dropped by 23,000 in July, while May and June employment gains were revised lower by a combined 103,000 jobs. The economy averaged only 34,000 new jobs per month over the previous 12 months.
Yet the unemployment rate hasn’t surged in tandem with those numbers.
According to Trading Economics, the unemployment rate stood at 4.3% in January, with roughly 7.4 million Americans unemployed. By July, unemployment dropped to 4.1%, with the number of unemployed Americans standing at 6.9 million.
Still, there is an important caveat.
Since January, the worker participation rate dropped by 0.7 points, while the employment-population ratio declined by 0.5 points. In July alone, Reuters reported, another 264,000 people left the labor force.
So one of the reasons why unemployment remained low was simply that fewer Americans are participating in the labor market.
Nonetheless, the AI-driven layoff story hasn’t exactly played out the way many expected.
Is AI hurting some workers while creating jobs elsewhere?
BofA’s analysis shows that the impact of AI is a lot more uneven than catastrophic. Young workers, though, are getting the short end of the stick.
BofA points to growing unemployment among people aged 22 to 27, particularly recent college graduates.
Though tariff uncertainty was initially touted as the core reason for hiring postponements, the numbers among recent graduates haven’t improved since.
Two white-collar areas in particular were earmarked by BofA.
Information technology boasts the highest AI usage in BofA’s dataset, at 42.1%, and a 1.9% decline in labor demand between January and June. Also, finance and insurance, where AI usage is at 34.8%, experienced a 1.1% decline.
At the same time, though, AI is simultaneously creating new demand somewhere very different.
The tremendous amount of capital required to build data centers is flowing into construction, electrical equipment, power infrastructure, and manufacturing. In fact, according to the bank’s estimates, nonresidential construction added 95,000 jobs year to date, while AI-related manufacturing industries added another 32,000.
Collectively these two categories represented 25% of new private-sector jobs created this year, according to BofA.
What does Bank of America’s AI jobs call mean for investors?
For AI stock investors, BofA’s findings are perhaps more significant than a simple debate over unemployment.
The first implication is support for AI capex, which has been perhaps the biggest debate over the past several months.
Naturally, one of the big bear cases surrounding AI is that companies will face political and economic resistance if we see mass unemployment due to the technology. So far, Bofa’s research team sees little evidence of that happening, and it observed that spending is actually translating into a meaningful source of employment in construction and manufacturing.
That also favors businesses that are positioned closest to the infrastructure buildout. If businesses can continue growing their AI capacity without materially impacting employment numbers and consumer demand, the investment cycle has remarkably more room to run.
But there is a second, less obvious implication.
If the labor market stays resilient, we would continue to see wage pressure and firmer economic activity, which prevents a drop in interest rates. Lower interest rates are an important consideration for richly valued growth stocks whose valuations depend on future earnings.
Also, if AI is more about augmenting workers than cutting labor costs, investors will demand more evidence that the spending is generating enough incremental sales and productivity to justify those investments.
For now, the scenario is mostly favorable.
AI spending continues to rise, infrastructure investment is creating new jobs, and widespread tech unemployment remains mostly absent. For AI investors, that is likely to extend the runway for the boom.