Copy article

Goldman-Sachs: New Report Finds 20% AI Job Displacement Dwarfs 8% Job Augmentation

ended 14. April 2026

Epoch AI and Ipsos found 20% of US full-time workers say AI has taken over parts of their job. Only 15% said they had started doing new tasks at work that they wouldn't have done without AI. 

Goldman Sachs estimates AI substitution is wiping out roughly 25,000 jobs per month, while augmentation adds back about 9,000, creating a net loss of 16,000 monthly, with the pain falling hardest on Gen Z and entry-level workers.

The route to AI implementation is still bumpy because of the problems that come with relying on probabalistic systems that are happy to deliberately conceal their inner workings and motivations.

  • Klarna replaced roughly 700 customer service agents with AI, watched quality deteriorate, and began rehiring'
  • McKinsey's AI platform was breached in two hours via SQL injection exposing 46.5 million chat messages. 

Is AI showing its colours as “the cmperor's new clothes?” in the consultancy space?

We'd like your views:

  • Displacement is outpacing augmentation. Is "AI makes workers more productive" still credible. or is it a bare-faced lie from boards?
  • Entry-level workers are absorbing the sharpest job losses. What obligations do employers and the state have to people whose roles are automated before reskilling programmes exist?
  • If the firms advising on AI augmentation and replacement can't secure their own platforms, what should clients conclude?
  • Klarna's AI replacement failed publicly. How many similar reversals are happening quietly, and what's the real cost of "move fast, fix later" workforce decisions?

2 responses from the Newspage community

Copy all

Star Quote
Copy

What is clear from this report and the fallout for these sort of corporate decisions and strategies is the pattern is consistent: AI removes work before organisations have the governance or honesty to deal with what follows.

The question isn't which pathway firms choose. It's who carries the cost while they're deciding. it's disappointing to see that when it comes to the boardroom, hype seems to have more value than humanity in the workplace.
Copy

Goldman Sachs puts the net monthly job loss at 16,000. That is not a productivity story. It is a structural one. The argument that AI makes workers more productive was always contingent on system redesign, not substitution. Most organisations skipped that step.
Entry-level roles weren’t just cheap labour. They were how organisations built capability over time. Remove them without a redesigned entry point and you don’t just lose headcount — you lose the pipeline. Organisations are teaching people to use AI. Almost none are redesigning the work itself. That is a different problem, and a harder one.
The more important question is not which firms reversed course. It is how many absorbed the cost quietly — in quality loss, customer churn, and rehire cycles that never made the announcement.
The consultancy breach question is fair but narrow. The larger issue is that firms advising on AI transformation are operating ahead of their own governance maturity.