Goldman Sachs CEO Solomon writes in NYT: AI doomsday panic is exaggerated, white-collar employment will take the biggest hit but entirely new positions will emerge

Goldman Sachs CEO David Solomon published an opinion piece in The New York Times on Tuesday, stating bluntly that the panic about AI triggering a "job apocalypse and mass unemployment" has been seriously exaggerated. Solomon argues that AI will not eliminate jobs on a catastrophic scale, but will instead boost worker productivity, shift employees toward higher value-added tasks, and create entirely new positions centered around the management, implementation, validation, and regulation of AI systems. However, he also acknowledges that white-collar industries such as banking, law, accounting, software development, and customer service will face the greatest impact. (Background: Charging $25,000 per day, two former fund managers conquer Wall Street through AI financial training) (Context: The Wall Street Journal slams stablecoins as "private money": posing major economic risks) Will artificial intelligence trigger mass unemployment? This central question troubling global labor markets has recently received one of Wall Street's most weighty voices. Goldman Sachs CEO David Solomon published an opinion piece in The New York Times on May 26 titled "I'm the C.E.O. of Goldman Sachs. The A.I. Job Apocalypse Is Overblown.", systematically rebutting the pessimistic narrative that AI will eliminate human jobs. In the article, Solomon explicitly points out that while AI will indeed bring impact to specific industries, historical experience has repeatedly proven that technological progress will ultimately create more jobs than it destroys. His argument rests on three core pillars: the historical trajectory of productivity gains, the chain effects of hyperscale capital expenditure, and the natural emergence of new job categories. Solomon acknowledges that the rapid development of AI has indeed sparked widespread labor market anxiety. Goldman Sachs economists predict that within the next decade, AI could automate approximately 25% of current work hours, with white-collar-intensive industries such as banking, law, accounting, software development, and customer service being most significantly affected. The work in these industries involves substantial data processing, document review, code writing, and standardized communication—precisely the areas where current AI systems excel. But he also emphasizes that the U.S. economy has experienced similar situations many times in the past, from the massive transfer of agricultural labor during the Industrial Revolution to the structural unemployment caused by manufacturing automation in the Information Age. Each technological shock was ultimately successfully absorbed by the economic system, with overall employment rates and living standards continuing to rise. AI will likely repeat this path: while eliminating some positions, it will expand other positions on a larger scale. Solomon points to a specific and quantifiable piece of evidence: in this year alone, hyperscalers plan to invest approximately $700 billion in capital expenditure (capex), and this massive investment has already driven a surge in employment in the U.S. construction industry. The construction of AI infrastructure such as data centers, fabs, and fiber optic networks requires substantial labor, from rebar workers to electrical technicians—positions that cannot be replaced by AI but are emerging in large numbers due to AI's development. This phenomenon reveals a key dimension overlooked by the "job apocalypse" narrative: AI is not just a technology that replaces jobs, but also an engine driving real economy investment. Behind every new data center and every new high-performance computing production line are thousands of construction and operations positions. Although the overall narrative leans optimistic, Solomon does not shy away from the real impact AI brings to the labor market. He specifically points out that white-collar industries will be the "epicenter" of this wave of technological change: In the banking industry, AI is already capable of handling loan approvals, risk assessment, and regulatory compliance—work traditionally done by analysts; in the legal industry, AI document review systems can complete document comparison work in seconds that previously required dozens of paralegals weeks to finish; the level of automation in accounting and auditing is also accelerating. However, Solomon believes these industries will not disappear as a result, but will undergo a qualitative transformation in job structure. Low value-added repetitive work will be replaced by AI, but the construction, management, validation, and regulation surrounding AI systems will generate a large number of entirely new professional positions. These new positions require higher judgment, more complex problem-solving abilities, and the capacity for critical evaluation of AI system outputs. In the article, Solomon mentions that AI will give rise to at least four emerging job categories: AI system administrators, responsible for daily operations and exception handling; AI implementation consultants, helping enterprises integrate AI into existing processes; AI validation engineers, ensuring the accuracy and fairness of AI outputs; and AI compliance specialists, handling regulatory and ethical issues. These positions are not science-fiction-like visions of the future, but are already gradually taking shape at Goldman Sachs and other Wall Street institutions. It is worth noting that Solomon is not speaking from an ivory tower. Goldman Sachs itself has already deployed AI tools extensively internally, from auxiliary analysis of trading strategies to intelligent response systems for customer service. The experience he shares in the article comes precisely from this Wall Street giant's firsthand observations during its AI transformation process.