More and more business leaders are pushing back on the claims of massive AI-induced job losses.
Goldman Sachs CEO David Solomon has joined the growing chorus of executives who believe the so-called AI job apocalypse is overblown. Writing in The New York Times, Solomon acknowledged that AI will unquestionably reshape the labor market — but argued that fears of mass unemployment fundamentally misread how economies adapt to technological disruption.
The 25% Number — And What It Actually Means
Goldman Sachs’s own economists estimate that AI may automate 25% of current work hours over the next decade. That sounds alarming. But Solomon’s point is that automating work hours is not the same as eliminating jobs.
His argument: when people are freed from routine tasks, they don’t sit idle — they take on more complex, higher-value work. He offers a sharp illustration from his own career: as a first-year banking analyst, building a simple stock performance graph took six hours of combing through microfiche archives of the Wall Street Journal. Today, it takes seconds. And Goldman hasn’t hired fewer people — it has hired more.
“With more sophisticated tools, the complexity of our work naturally expands,” Solomon writes. “Do any of us feel like we have less to do these days despite the convenience of Excel, email or Zoom?”
This mirrors the view of Scale AI CEO Alexandr Wang, who has argued that AI won’t eliminate roles so much as transform them — with every employee effectively becoming a manager overseeing a pod of AI agents. Scale AI’s interim CEO Jason Droege has also called the near-term job elimination narrative “very overhyped.”
The Counterargument Is Real
Solomon is careful not to dismiss the disruption. A Stanford study cited in his piece found that in the occupations most exposed to automation — software engineering, customer service — entry-level employment has already fallen 16% relative to less-exposed roles. White-collar workers, including accountants, bankers, and lawyers, will see many of their tasks automated.
The pessimist camp has powerful voices. Anthropic CEO Dario Amodei has warned of very high unemployment and inequality as a direct consequence of AI, with corporate America potentially replacing up to 3 million workers by end of 2026. Amazon CEO Andy Jassy has gone further than most corporate leaders in stating plainly that Amazon’s overall corporate headcount will shrink due to AI — a signal that new AI-created roles may not fully offset those lost.
Meanwhile, finance job openings are already at their lowest since the financial crisis — a sector that sits squarely in the crosshairs. Sam Altman and Perplexity’s Aravind Srinivas have both acknowledged job losses are coming, even if they believe better jobs will eventually replace them.
Three Reasons Solomon Is Still Optimistic
Solomon leans on three core arguments to make his case:
1. History keeps proving the apocalypse wrong. From electrification in the 1900s to the digital revolution of the 1990s, the US economy has consistently created new jobs in response to disruption. Since 1962, civilian employment has grown roughly 145%, outpacing population growth of 128%. The health care industry alone now employs over 18 million — a category that barely existed at scale a century ago.
2. Technology doesn’t automatically eliminate demand for human connection. ATMs didn’t kill banking jobs. The internet didn’t put real estate agents out of business. Television didn’t kill live entertainment. Solomon’s argument is that technology tends to highlight — not replace — the value of human judgment and relationships.
3. The American labor market is unusually dynamic. US companies destroy and create between 25 to 35 million jobs annually. AI will accelerate this churn, not end it. New roles are already emerging — managing agentic AI systems, validating AI outputs, compliance, and workflow integration — all of which require human oversight.
The New Jobs Being Created
Solomon points to a concrete example: Goldman Sachs’s economists estimate that the surge in data center construction driven by AI has created over 200,000 construction jobs since 2022. While AI may reduce headcount in regulatory reporting or client onboarding at Goldman itself, it frees the firm to hire more bankers, traders, and asset managers.
The Naukri.com CEO echoes this from the hiring side, noting that while AI is automating certain roles, new jobs are visibly emerging — particularly for those who build complementary AI skills rather than compete with AI directly.
What Needs To Happen
Solomon doesn’t pretend the transition will be painless, and calls for a joint public-private response: government investment in vocational schools and community colleges, and private sector commitment to reskilling workers and reimagining on-the-job training.
The historical lesson he draws from is John Maynard Keynes’s famous 1930 prediction that people would be working just 15 hours a week by 2030. That forecast never materialized — not because productivity stalled, but because human ambition kept finding new things to do with the extra capacity.
Solomon’s bet is that the same dynamic holds now. The job apocalypse narrative, he argues, systematically underestimates AI’s potential to spur a productivity revival — one that creates demand for human work, rather than extinguishing it.