Goldman Sachs economist Joseph Briggs estimates that AI could displace up to 9% of US workers and expects most of the job losses to arrive in a recession, with hiring already running about 10,000 below trend in the most AI-exposed industries. Separately, Gallup finds that people who use AI weekly are more likely to hold a quality job than people who never use it, 52% to 32%, and Amazon and New York each announced entry-level hiring and training commitments this week.
By the Numbers
Up to 9%: share of US workers Goldman Sachs says AI could displace during the transition, with slower hiring cutting employment another 3% (Goldman Sachs via Fortune).
194,000: initial jobless claims for the week ending October 17, down 6,000 (U.S. Department of Labor).
52% vs. 32%: share of weekly AI users and of non-users who hold a quality job (Gallup).
30,000: entry-level UK roles Amazon pledges by 2030, matching the 30,000 corporate jobs it cut over the past year (Fortune).
40,000: New Yorkers the AI Jobs Institute aims to train by 2032 on $10 million in public money (Fortune).
Layoffs and Company Decisions
Goldman Sachs says AI could displace up to 9% of workers, and layoffs will likely wait for a recession
Goldman Sachs economist Joseph Briggs published a note on October 22 estimating that up to 9% of US workers could be displaced during the AI transition, assuming AI eventually lifts productivity 15% in the US and other developed economies. Slower hiring could cut employment another 3%. Briggs reviewed 25 occupations hit by earlier technologies: separation rates rose in 19 of them, and employment fell about 25% over five years on average, which he calls likely upper bounds for AI. Routine jobs held steady in expansions and dropped sharply in the downturns of the 1980s, 2001, 2008, and 2020. Today only 23% of US establishments have built AI into regular workflows, and job growth in the most AI-exposed industries runs about 10,000 below trend over three months, roughly offset by AI-related construction hiring. Briggs wrote that “the timing and extent of automation and labor disruption will likely track the overall economic cycle.” Goldman is an investment bank, and Fortune does not link the note itself.
Fortune, Nick Lichtenberg, October 22, 2026. Read the report
Why it matters: Goldman places the main layoff risk in the next downturn, which gives HR leaders time before cyclical cuts and automation arrive together. Build redeployment and retraining capacity now.
Amazon pledges 30,000 UK entry-level roles after cutting 30,000 corporate jobs
Amazon announced its Youth Promise on October 20: 30,000 new entry-level jobs in the UK by 2030, aimed at young workers and at the roughly one million people in the UK who are not in education, employment, or training. Starting salary is £31,000 (about $41,500). Amazon cut 30,000 corporate jobs over the past year to restructure around AI, and let go about 1,000 white-collar workers in India, the UK, and its retail business in early October, according to Fortune and HR Brew. Microsoft, Cisco, and IBM have made similar pledges through the UK government’s Youth Guarantee. UK country manager John Boumphrey said, “When I visit our sites, I meet young people who had no idea where to start six months ago.” The report does not say which functions the new roles sit in or how AI changes them.
Fortune (via HR Brew), Kristen Parisi, October 23, 2026. Read the report
Why it matters: The pledge and the cuts sit in different countries and at different job levels, so they do not offset each other. Ask which roles carry the hiring and which carry the cuts.
Initial jobless claims fall to 194,000
The U.S. Department of Labor reported on October 22 that initial claims for the week ending October 17 were 194,000, down 6,000 from a revised 200,000. The four-week average is 197,750. Insured unemployment for the week ending October 10 was 1,728,000, up 25,000, and the insured unemployment rate held at 1.1%.
U.S. Department of Labor, release USDL 26-1613-NAT, October 22, 2026. Read the report
Why it matters: Few laid-off workers are filing for benefits, which fits Goldman’s expectation that AI cuts accelerate only in a downturn. Treat today’s claims level as the baseline a recession would break.
Reskilling and Education
New York commits $10 million to train 40,000 people for AI jobs
Gov. Kathy Hochul announced the AI Jobs Institute on October 22, a hub in Long Island City, Queens, designed and run by the nonprofit Pursuit. The goal is to train up to 40,000 New Yorkers by 2032, connect them with employers, and produce workforce policy research. City and state put in $10 million; private funders include Ascendium Education Group, Google.org, Wells Fargo, Con Edison, and Adobe. OpenAI, Anthropic, and Google supply credits, certificates, and platform access, and sit on the advisory council. Each builds the tools trainees will learn. Pursuit runs a 630-hour AI course that about 400 to 500 people have completed, including shorter workshops. By Pursuit’s account, 60% of its participants lack a four-year degree and half receive public assistance. The state’s release sets no placement or wage targets.
Fortune, Catherina Gioino, October 22, 2026. Read the report
Why it matters: A 40,000-person target over six years comes with no placement or wage goals, so the first measure will be enrollment. Employers considering hires from the program should ask Pursuit for placement and wage outcomes by cohort.
Gallup: 41% of US workers hold quality jobs, and weekly AI users are more likely to
Gallup, with Jobs for the Future, the Families and Workers Fund, and the W.E. Upjohn Institute, released its second American Job Quality Study. It counts 41% of US workers in quality jobs, unchanged from 2025, using five dimensions: fair pay and benefits, a safe and respectful workplace, growth opportunities, a voice in decisions, and sustainable schedules. Among employees who use AI at least weekly, 52% hold quality jobs, against 32% of those who never use it. One in four employees reports discrimination or unfair treatment, and one in four lacks opportunities to advance. Axios reports that AI use is more common in the better jobs typically held by college graduates, and those workers are the ones who worry AI will replace them. The survey covers more than 17,000 workers and cannot show whether AI use leads to better jobs or the reverse.
Axios, Emily Peck, October 22, 2026. Gallup report date not stated on the study page. Read the report
Why it matters: Weekly AI use clusters in better jobs, so training offered only to current high performers widens a gap that already exists. Check who in your organization gets AI tools first.
What Workforce Leaders Are Watching
If layoffs track the business cycle, as Goldman expects, which retraining and redeployment programs will companies have running before the next downturn begins?
Amazon cuy 30,000 corporate jobs and is hiring 30,000 UK entry-level workers: do the new roles lead into the jobs that remain after AI restructuring?
Weekly AI users hold quality jobs at 52% against 32% for non-users: who in your organization gets AI tools and training first, and who waits?
New York’s 40,000-person training target has no placement goals: what outcomes would your hiring teams need published before they recruit from it?
This briefing was prepared automatically by the Workforce Rewired research assistant. All stories include direct source links.
A firm-level data point that fits Briggs' slower-hiring channel. I read 4,335 earnings calls from S&P 500 companies outside tech since January 2024 and pulled every passage where an executive tied AI to the company's headcount. It happened on 41 calls, about one in a hundred, though the share has nearly tripled since 2024.
On 22 of those 41 calls the claim was that AI lets the company grow without adding people, and on 5 more that it covers for people who leave. Regions' CFO described attrition "that we won't have to replace". Only 14 described cutting roles. So far executives describe roughly what Goldman expects before a downturn: fewer hires rather than layoffs.
Write-up with each quote linked to its call (I work at Equibles, where it's published): https://equibles.com/research/is-ai-cutting-jobs-outside-tech-evidence-from-4-335-s-p-500-earnings-calls
For the HR people here: does that match what you see, with freezes in specific functions rather than cuts?