
"Goldman Sachs Group Inc. ( NYSE: GS) has made it plain that it will start to cull its workforce based on the ability of some artificial intelligence (AI) programs to do what humans do better than humans. However, the AI-based decisions are not entirely about AI advantages. The bank has always been known for ruthless management. It has asked executives to find "underperformers." This could affect about 1,000 people this year."
"A memo obtained by the media showed, however, that the Goldman decision was not just about poor performance among certain workers. "The rapidly accelerating advancements in AI can unlock significant productivity gains for us," a management memo read. Many of the people who lose jobs to AI will not get those jobs back, at Goldman, or anywhere else in the industry. The AI contribution to investment banking cost cuts is permanent, particularly in several categories."
"The most expensive jobs to be replaced are those of analysts who have recently graduated, usually from business schools. In their first two or three years, these people work extraordinarily long hours. Their reward is pay packages in the low hundreds of thousands of dollars. Their primary job is analysis of huge amounts of data, to find key edges in trading methods, corporate performance, and patterns for macroeconomic investments."
Goldman Sachs plans to reduce its workforce by evaluating which roles can be performed better by artificial intelligence, with management asking executives to identify underperformers. About 1,000 employees could be affected this year, and it is unclear if hiring will offset the cuts. Management cited rapidly accelerating AI advancements that can unlock significant productivity gains. Many displaced workers are unlikely to regain similar positions within Goldman or the industry, as AI delivers permanent cost reductions in several areas. Jobs at greatest risk include junior analysts who process large data sets, back-office data processors, and human-resources roles.
Read at 24/7 Wall St.
Unable to calculate read time
Collection
[
|
...
]