The Federal Reserve Bank of New York recently warned that the unemployment situation for recent college graduates has “deteriorated noticeably.” Though there are multiple factors contributing to this shift, including President Trump’s economic policies, the adoption of AI in technical fields like finance and computer science is playing a role. These fields have been targeted for disruption because they are more easily automated - and they tend to have the highest labor costs. As AI matures, these job trends will likely spread to other industries.
But is this really all about AI?
Anthropic released its first Economic Index paper with analysis based on data from Claude usage. It reveals a few noteworthy trends, none of which seem to suggest large-scale replacement of workers:
AI use is concentrated in software development and technical writing tasks.
36% of occupations see AI use in at least a quarter of their tasks.
4% of occupations use AI across three quarters of their tasks.
AI use leans more toward augmentation of human capabilities compared to automation.
AI use is more prevalent for tasks associated with mid-to-high wage occupations. It’s lowest for both the lowest and highest paid roles.
Despite these muted trends, Anthropic’s CEO, Dario Amodei, issued a stark prediction that AI could automate away up to 50 percent of all entry-level white-collar jobs within five years. But Allison Morrow suggests taking this fearmongering with a grain of salt - especially when it comes from the very people who stand to profit from the perception that this technology is reaching a point of approximating human thinking. Morrow’s warning rings true given the MIT study finding that 95% of generative AI pilots failed to achieve rapid revenue acceleration for businesses and delivered little to no measurable impact.
The tech labor market story is more complicated than just an AI takeover. A bigger driver, SignalFire argues, is the end of the low interest rate period from 2020 to 2022 that led to over-hiring and inflation. Another culprit is a delayed change to a decades-old tax provision that was buried in the 2017 Tax Cuts and Jobs Act that has reportedly contributed to the loss of hundreds of thousands of high-paying tech jobs. The change to section 174 of the Tax Code, which allowed companies to deduct 100 percent of qualified R&D spending in the year they incurred the cost, directly incentivized the hiring of American tech workers. It is no coincidence that Meta, Microsoft, Alphabet, Amazon, Salesforce, and other tech companies announced large workforce cuts just as the change to section 174 was coming into effect. While publicly the companies blamed overspending and AI, their 10-Ks revealed the changes in their R&D expenses.
Companies may be adjusting their behavior in anticipation of more capable AI tools by underinvesting in job training, mentorship, and other programs aimed at entry-level workers. SignalFire’s latest jobs report reveals that entry-level hiring is collapsing in tech. As AI tools take over more routine, entry-level tasks, early-career talent has a harder time breaking in. They warn that while AI might reduce short-term need for junior hires, skipping them entirely risks breaking the long-term talent pipeline.
Questions to consider
For companies citing AI adoption as the driving force for reducing headcount, how are they measuring the effectiveness of automation efforts? What are their plans for maintaining their long-term talent pipeline?
