Report: Generative AI Could Automate 25% of US Jobs

Advances in artificial intelligence (AI) could eventually impact 300 million jobs worldwide.

That’s according to new research released Monday (Mar. 27)  by Goldman Sachs, which finds that generative AI technology such as ChatGPT could lead to a boom in productivity, but also create significant disruption for the labor market.

The findings — as reported by the Financial Times and Mint — say that automation could impact 25% of jobs in the U.S. in Europe, with attorneys and administrative workers at the highest risk of job loss.

According to the report, about two-thirds of jobs in Europe in the U.S. are exposed to some level of AI automation. Most workers would see less than half their tasks automated and could probably keep their jobs, with time free for other work-related activities.

The report posits that this should apply to 63% of U.S. workers, while another 30% has physical or outdoor jobs that AI likely won’t touch. That last 7% are in roles where AI could automate at least half of their jobs, and thus replace them.

Goldman’s report follows recent research from OpenAI that estimates that generative pre-trained transformer (GPT) models, and software tools built atop them, could affect up to half of the tasks necessary for 19% of the jobs in the U.S..

“Our analysis indicates that the impacts of LLMs like GPT-4, are likely to be pervasive,” the study says, adding that 80% U.S. workers could see at a tenth of their tasks affected in “some way” by ChatGPT.

Jobs that already employ software-driven tasks could potentially face greater disruption from AI tools, while industries that use a lot of manual labor, including food service and hospitality, forestry and others, will likely see the least impact, the report said.

PYMNTS looked at the effect AI could have on the workforce in a recent conversation with Matthew Tillman, CEO at automated accounts payable solution OpenEnvoy.

“[AI] no longer just has an implication as a use case, it has an implication on their business,” Tillman told PYMNTS CEO Karen Webster.

The implication, he added, is that businesses can see right away how many — or how few —  people they might need to man a department in the near future, leading to a world where more CFOs think about “how can we run a department of one?”

“The AI technology develops the script that a business would’ve put toward low-cost labor. And it takes action on that script,” Tillman said.

In fact, in the generalized finance space, areas like bookkeeping and non-tax audits can “go away today,” he added. In Tillman’s mind, a business with a dozen people managing $6 billion worth of spend is one with 11 too many people.