Goldman Sachs says AI could replace 300 million jobs

White-collar workers are just getting used to having ChatGPT polish their emails, and Wall Street is already calculating how many of them can be phased out.

A report from Goldman Sachs warned that generative AI could replace the equivalent of 300 million full-time jobs globally. In the US and Europe alone, roughly a quarter of all current work tasks could be fully automated by AI.

White-collar work is on the chopping block

Unlike previous waves of automation that replaced factory workers and manual laborers, generative AI flips the script entirely.

According to Goldman Sachs, the administrative sector faces the highest exposure, with 46% of office tasks ripe for automation. The legal field follows closely behind at 44%.

Physical jobs, on the other hand, are largely insulated for now. The report estimates that only 6% of tasks in construction and a mere 4% in maintenance could be handled by AI.

Governments are already trying to ease public anxiety. UK Technology Secretary Michelle Donelan told The Sun that the goal is ensuring AI complements workers rather than disrupts them, making jobs better instead of taking them away.

The real threat isn't layoffs, it's lower wages

While eye-popping layoff numbers grab headlines, some labor researchers argue that a quieter, more painful shift is already underway: wage cuts.

Carl Benedikt Frey, director of the future of work at the Oxford Martin School, told BBC News that precise job loss predictions are impossible to make. Instead, he points to market dilution. ChatGPT essentially allows anyone with average writing skills to generate polished prose, meaning journalists and copywriters will face a flood of new competition that naturally drags down pay.

Frey drew a parallel to the rise of GPS and Uber. Knowing every street in London used to be a prized, lucrative skill. Once GPS arrived, that advantage evaporated overnight, driving wages down by about 10% for incumbent drivers. It didn't eliminate drivers; it just made them cheaper.

Goldman Sachs tried to offer a silver lining, noting that 60% of today's workforce is employed in roles that did not exist in 1940, and predicting that generative AI could eventually lift global GDP by 7%. But the report also cited counter-research showing that technological shifts since the 1980s have eliminated jobs faster than they created new ones.

Torsten Bell, chief executive of the Resolution Foundation think tank, cautioned against taking doomsday forecasts as gospel. He noted that all firm predictions should be taken with a very large pinch of salt, emphasizing that we still have no idea how quickly businesses will actually adopt these tools.

Why it matters

We think workers shouldn't panic over the 300 million headline, but ignoring the warning signs would be a mistake.

If your daily grind revolves around summarizing documents, drafting administrative reports, or basic legal research, your competitive moat is evaporating fast. The practical takeaway here isn't to look for a completely new career tomorrow, but to start leveraging generative AI immediately. You need to automate the 40% of mundane tasks yourself before your employer realizes an algorithm can do it without you.

For companies and the broader market, this trend threatens entry-level hiring. If AI handles junior administrative and drafting duties, the traditional apprenticeship model breaks down, leaving firms with fewer paths to train senior talent.

Goldman's forecast of a 7% bump in global output sounds great on paper, but the distribution of those gains is the real sticking point. As academic researchers pointed out, the immediate pressure will be on wages, not employment numbers. Having more workers chasing cheaper tasks is a recipe for wage compression, and that is a problem productivity gains alone cannot solve.