Goldman Sachs executive: Millions will lose their jobs, but will find another

Joseph Briggs, head of the global economy team at Goldman Sachs’ research department, said in the latest episode of the bank’s “Exchanges” podcast that he expects about 9% of the US workforce to be out of a job following the adoption of AI throughout the economy.

● What really resulted in the rejection of the Anthropic and OpenAI IPOs
● Opens new shorts on AI: Michael Berry warns of “the beginning of the end”
● On the way to $5 million in revenue per employee: the two clocks CEOs need to turn on

“Ejection of 9% of the workers due to artificial intelligence equals about 15 million workers”. Briggs said, comparing the scale of change to the technology upheaval of the late 1990s and early 2000s. These workers, according to him, will be forced to leave their current positions and find new jobs.

The effects are already visible on the ground

In sectors where AI tools are already being used, such as high-tech, management consulting and graphic design, Briggs estimates that the technology cuts between 10 and 15 thousand jobs from the monthly employment growth rate.

However, Briggs rejected the claim by many technology leaders that jobs will disappear forever, arguing that this approach only focuses on the jobs that are being destroyed and ignores those that are being created.

History on his side, he said: “If we look back over the last 80 years, about 85% of job growth was driven by the creation of new positions as a result of technological changes.”

According to him, the labor market is still in constant motion, with around 30 million jobs being created and around 29 million being deleted every year. Against this background, even a mere 5% acceleration in the rate of job creation would be enough to re-absorb all those that AI would throw out.

Neil Thompson of the Massachusetts Institute of Technology (MIT), who was also a guest on the podcast, argued that change will be slower than the rapid rate of improvement in AI capabilities suggests.

Technological capability is only the first step, he explained: an AI system also needs access to the right information—a complex task in fields like medicine, where privacy laws are an obstacle—and it must be cheap enough to make it worthwhile to operate. These hurdles mean that the pace of technology adoption may lag significantly behind what AI is capable of doing on a technical level.

Most jobs can be partially automated rather than completely eliminated, Thompson said, and the outcome depends on the specific tasks the machines will take on. When GPS automated taxi drivers’ professional knowledge of city routes, wages did drop, but the number of drivers jumped dramatically.

The way he sees it, AI is like a “rising tide” that workers can foresee and prepare for, rather than a “crashing wave” that washes over them.