Goldman: AI layoffs prolong unemployment, reduce older workers' pay

By Sophie Lindqvist

Workers displaced by AI-related layoffs last year took an average of one month longer to find new roles than workers did a decade earlier, according to a Goldman Sachs study whose published summary does not identify the geography.

The study’s central claim, attributed to Goldman Sachs economist Pierfrancesco Mei, is that AI’s main disruption now appears in re-employment rather than a new wave of immediate layoffs. Mei compared the phase with earlier technology shifts, including automation and personal computing. That comparison is a historical analogy, not proof that AI caused each layoff or outcome.

The accompanying commentary describes a booming economy alongside widespread layoffs and unresponsive job applications, then says those conditions discourage workers from leaving jobs. It supplies no GDP, layoff or application-response figures, so that context remains a description rather than a measurable baseline.

What changes after an AI-related layoff?

Four reported outcomes translate the study into a weaker case for rapid recovery at the same pay:

• The search after an AI-related layoff lasted one month longer than the comparable search a decade earlier.
• About 10% of affected workers were expected to remain unemployed after one year.
• Three-quarters of laid-off workers who found jobs entered roles with similar or greater AI exposure than their previous positions.
• Among workers over 55 who found new jobs, the typical outcome was a 10% pay cut from the previous salary.

For an affected worker, the Goldman figures imply a longer interval without earnings, not merely a changed title. The 75% finding means that finding comparable work may still mean carrying similar or greater exposure to the technology associated with the displacement.

It does not show that every displaced worker moved into the same occupation. The 10% pay reduction applies specifically to the over-55 group and cannot be generalized to every age group.

What do the graduate figures show?

At the population level, recent college graduates faced a 5.4% unemployment rate in the last quarter, compared with 4% for the overall population. The 1.4 percentage-point gap makes the graduate rate 35% higher than the population comparator, a relevant benchmark for new entrants evaluating available openings rather than relying only on the headline rate.

Axios supplied the underemployment measure: “about 42% of employed recent graduates held jobs that don’t require a degree.” The article presents it as evidence that a title does not guarantee degree-matched work. It does not establish that AI caused the graduate unemployment rate or the degree mismatch, and it provides no breakdown by field, age or occupation.

What does the summary leave unresolved?

The Goldman summary reports no sample size, country, exact quarter, comparison methodology, confidence interval or revision status for the layoff figures. “Last year” and “last quarter” lack calendar dates.

No industry or occupation breakdown appears to rank sector exposure. Without those details, readers cannot treat the figures as national estimates, compare them month by month or infer causation from the AI label.

Could AI create new roles?

The source says AI could eventually produce a new subset of jobs, but those openings may require workers to upskill first. It gives no number, timetable, occupation list or qualification standard, so the promise remains conditional.

For a displaced worker, the source offers no way to match future hiring to a specific course or credential. It documents present re-employment costs, not a forecast of guaranteed openings.

What does the sponsored Symphony offer?

Separate from the Goldman research, the article carries a Black Friday promotion presented through Wix for Symphony. The promotion calls Black Friday “a boom time for sales and the hardest shopping season of the year.” It pitches Symphony as a team of AI agents for independent business owners, not a labor-market program.

The described agents build a Black Friday offer with discount codes, gift badges and dates that switch on and off automatically. Other agents reach customers and recover abandoned carts through three reminders that run around the clock and stop when a customer buys.

An overnight brief is supposed to show what arrived, what needs shipping and what is waiting, rather than leaving the owner with a backlog. The promotion says Symphony Pro and above, especially annual plans, give the AI team “more room to work.”

The sponsored email was written by David Vendrell, edited by Nick Comney, polled and copy-edited by Kait Cunniff, and published by Darline Salazar. Those product claims sit outside Goldman Sachs research, and the source provides no independent performance figures to test the agents’ reliability or commercial results.