Behind the Averages

Defining pattern

This week’s numbers looked steady on the surface and uneven underneath. Pew found that employment in the jobs most exposed to AI has grown, while women without four-year degrees in those same jobs lost ground. Barclays and American Express each announced cuts of about 5% while profits rose, as eBay, Flutter, and Qualcomm showed that not every layoff traces back to AI. Goldman Sachs argued that AI displacement will show up mainly in the next downturn, not in today’s low layoff numbers, and a new Senate bill sketched what a retraining safety net could look like. Federal and state training money moved in New York, and Monroe Community College won a new federal award for optics training. The common thread: totals and averages are hiding who carries the change, and when.

011. AI-Exposed Jobs Are Growing. For Women Without Degrees, the Pattern Runs the Other Way.

Pew Research Center published an analysis on October 22 comparing employment in 480 occupations in 2017 to 2019 with January 2023 through June 2026, using Current Population Survey data and excluding the pandemic years. Each occupation received an AI-exposure score from 0 to 3.5, based on Department of Labor O*NET data on whether generative AI can perform or assist with its tasks. Pew then compared men’s and women’s employment growth within each occupation. Overall, employment in the most AI-exposed occupations grew rather than shrank. But for each one-point increase in exposure, men’s employment grew 6.3% while women’s grew 0.2%, and the gap widened step by step from the least to the most exposed jobs.

Education sharpened the difference. Among workers with a high school diploma or less, each one-point increase in exposure was associated with 8.8% employment growth for men and a 9.5% decline for women. Among workers with some college, including associate degrees, the figures were 3.6% growth for men and a 6% decline for women. Among bachelor’s degree holders, neither group’s employment changed significantly as exposure rose. Pew notes that women without degrees often work in secretarial, receptionist, and customer service roles, which sit higher on the exposure scale, while men without degrees mostly work in low-exposure jobs like truck driving and construction labor, so the men’s figure applies to a small share of them. Pew cautions that the relationship “is not necessarily direct,” that other changes during this period may have affected employment, and that it is too early to tell whether exposure will lead to job loss or growth in the long run.

• Among workers with a high school diploma or less, each step up Pew’s exposure scale meant 8.8% more employment growth for men and 9.5% less for women.
• Secretarial, receptionist, and customer service roles, often held by women without degrees, rate relatively high on AI exposure.
• For bachelor’s degree holders, Pew found no significant gender difference as exposure rose.

“AI-exposed jobs are growing” and “some workers in AI-exposed jobs are losing ground” can both be true, and the second is easy to miss inside an aggregate. Pew doesn’t establish that AI caused the gap, but the pattern falls on a group workforce centers serve every day: women in administrative and front-line customer roles without a four-year degree. For workforce organizations, that argues for tracking placements and wages by gender and education, not just in total. For workers in these roles, it argues for building, and documenting, the judgment, relationship, and problem-solving parts of the job.

Impact: Emerging.

022. Barclays and American Express Cut About 5% of Staff as Profits Rise, With AI in the Explanation

Barclays said on October 22 that it is cutting about 5,000 jobs this year, roughly 5% of its workforce, as it uses technology such as AI to reduce fixed costs, according to Reuters. CFO Anna Cross said about 60% of the reductions fall on support functions, including developers, and that about 1,000 roles have already gone, with the rest due by year-end. Barclays reported third-quarter profit of £2.3 billion, up 12% from a year earlier, though below analysts’ forecasts.

A day later, American Express said it plans to eliminate about 5% of its global workforce, roughly 3,500 jobs, and recorded a $336 million restructuring charge while raising its full-year profit forecast. CFO Christophe Le Caillec said the changes would make the company “leaner, more agile, and also to prepare for AI,” according to Reuters. Neither company has said how many of the eliminated roles involve work AI is already doing.

• About 60% of Barclays’ cuts fall on support functions, including software developers.
• About 1,000 of Barclays’ 5,000 cuts have already happened, so roughly 4,000 remain this year.
• AmEx took a $336 million restructuring charge in the same quarter it raised its profit forecast.

These are cuts at profitable companies, and that changes what a layoff signals. AmEx’s phrase “prepare for AI” describes efficiency a company expects, not replacement it has demonstrated, and it’s worth reading that way. For workers in banking operations, support, and even software roles, strong earnings aren’t job security. For workforce organizations, financial-services back-office staff are a population to watch.

Impact: Immediate.

033. Not Every Layoff Is an AI Layoff: eBay, Flutter, and Qualcomm

eBay filed an 8-K on October 20 announcing plans to cut about 9% of its global workforce, with $130 million to $150 million in pre-tax charges, nearly all for severance. In the same filing, eBay said third-quarter results would come in above its earlier guidance. The filing does not mention AI; it says the plan aligns teams “with its top strategic priorities.” News reports estimate the cut at roughly 1,000 to 1,100 jobs, based on eBay’s last reported headcount. The restructuring is expected to be substantially complete by the end of the first quarter of 2027.

Also on October 20, Flutter Entertainment announced up to 1,400 job cuts across its UK and Ireland business, including hubs in Portugal and Romania, and attributed them to higher UK gaming taxes. Reuters reported the same day, based on an internal memo, that Qualcomm plans worldwide cuts after its annual performance reviews, with notifications starting in November and no number given. Counting is messy too: TechRepublic reported that the TrueUp tracker shows 196,349 tech workers affected by layoffs in 2026, while the GoLayoffs tracker shows about 65,000, because they define layoffs differently.

• eBay’s filing pairs a 9% workforce cut with better-than-expected results, and never mentions AI.
• Flutter attributes up to 1,400 cuts to higher UK gaming taxes, not technology.
• Two tech-layoff trackers put 2026 totals at about 196,000 and about 65,000.

With AI dominating the headlines, it’s tempting to file every layoff under it. The filings say otherwise: strategy, taxes, and performance reviews drove these. That matters for anyone designing a response. A worker cut because a function was automated needs different help than one cut because a company reshuffled priorities, and inflated “AI layoff” counts make it harder to see where retraining is actually needed.

Impact: Immediate.

044. Goldman: Up to 9% of Workers Could Be Displaced, Mostly When the Economy Turns

Goldman Sachs economist Joseph Briggs published a research note on October 22, reported by Fortune, estimating that up to 9% of workers could be displaced during the AI transition, assuming AI eventually raises productivity by 15%. Slower hiring would weigh on employment by another 3%. The 9% refers to gross job losses, not the unemployment rate, since some displaced workers leave the labor force. The figure has grown: Goldman estimated 6% to 7% in August 2025.

The new element is timing. Briggs found that the share of Americans in routine jobs has stayed stable during expansions and dropped sharply in downturns, and that “AI-related labor disruptions could well amplify the broader business cycle.” Across 25 occupations hit by past technologies, employment fell an average of 25% over five years once decline began, which he calls “likely upper bounds” for AI. For now, job growth in AI-exposed industries has run about 10,000 below its long-run trend over three months, roughly offset by AI-related construction hiring, and only 23% of U.S. establishments have built AI into regular workflows.

• Goldman’s estimate rose from 6% to 7% in August 2025 to up to 9% now, under the same 15% productivity assumption.
• Routine-job employment held steady in past expansions and fell sharply in downturns.
• Only 23% of U.S. establishments have built AI into regular workflows so far.

This is a projection, not a count, but the timing argument is useful. Companies under pressure in a downturn tend to make cuts they had been putting off, and automation gives them a reason. If that pattern holds, the period to build transition capacity, in rapid response teams, dislocated-worker funding, and training seats, is before the next recession, not during it.

Impact: Long-term.

055. Few Layoffs, Slow Hiring, and Paychecks Behind Prices

The Labor Department reported 194,000 seasonally adjusted initial unemployment claims for the week ending October 17, down 6,000. The four-week average fell to 197,750, which Reuters and CNBC reported as the lowest since 1969. Continuing claims, a rough gauge of how hard it is to get rehired, rose 25,000 to 1,728,000 for the week ending October 10. ADP’s preliminary estimate showed private employers adding an average of 18,500 jobs per week in the four weeks ending October 3, down from a revised 26,000.

Indeed’s Job Postings Index reached 103.8 on October 16, up 2.8% from a year earlier, and 68% of occupational sectors had more postings than a year ago. But new postings have “continued to drift sideways,” and Indeed said it “would be premature to say the labor market is strengthening.” Meanwhile, the Bureau of Labor Statistics reported median weekly earnings of $1,232 for full-time workers in the third quarter, up 1.5% over the year, while consumer prices rose 3.5%. Women’s median weekly earnings were $1,127, or 83.4% of men’s $1,352.

• The four-week average of initial claims, 197,750, is the lowest since 1969, while continuing claims rose to 1.73 million.
• Median weekly pay rose 1.5% over the year as consumer prices rose 3.5%.
• Women’s median weekly earnings were 83.4% of men’s in the third quarter.

This is the same low-fire, low-hire market Career Intel described on October 3, and the new data hasn’t changed the picture: keeping a job is easier than finding one. The wage numbers add a second squeeze, since a typical full-time paycheck is losing ground to prices. For anyone weighing an offer, compare it with living costs, not just a previous salary.

Impact: Immediate.

066. A Senate Bill Would Pay Unemployed Workers to Stay in Training

Sen. Ron Wyden, the ranking Democrat on the Senate Finance Committee, introduced the Stability and Opportunity for Workers Act on October 23. Its “Transition Wages” provision would provide 26 additional weeks of federally funded benefits for workers enrolled full time in approved training after they exhaust regular benefits, for up to 52 weeks in total, or 104 during high unemployment. Approved training includes programs offered by WIOA-eligible providers, Workforce Pell-eligible programs, and training for in-demand occupations. The bill would also require states to offer at least 26 weeks of regular benefits replacing 75% of prior earnings, subject to state maximums.

It would also create a $250-a-week Jobseeker Allowance for unemployed people without unemployment insurance, such as new labor-force entrants and self-employed workers, fund a tuition pilot for claimants in approved training, and rebuild Extended Benefits so that up to 52 extra weeks become available when unemployment reaches 8.5%, with full federal financing. The changes would take effect by January 1, 2028. CBS News framed the bill as a response to AI-related layoffs; it follows Sen. Mark Kelly’s Make AI Work for Americans Act, introduced last month, which proposes excise taxes on the AI industry to fund retraining.

• Transition Wages would cover up to 52 weeks of benefits for workers in full-time approved training, including WIOA-eligible programs.
• A $250-a-week Jobseeker Allowance would reach unemployed people who don’t qualify for regular benefits.
• Extended Benefits would scale to 52 weeks when unemployment reaches 8.5%, far above today’s 4.2% national rate.

Anyone who has worked with dislocated workers knows the usual reason people leave training: the benefits run out before the program does. Transition Wages targets that directly, and it uses the existing WIOA provider list rather than building a new one. It’s a proposal from the minority party, and passage is uncertain. But it’s a concrete marker of how an AI-era safety net might be built.

Impact: Long-term.

077. Pursuit Lands Federal and State Money in One Week. Upstate Access Is the Open Question.

The U.S. Department of Labor announced $95.5 million in Rapid Reskill Dislocated Worker Grants on October 19, covering 20 projects in 17 states. The grants aim to place dislocated workers in paid work within six months of enrollment, through full-time unsubsidized jobs, Registered Apprenticeships, or paid on-the-job training, in fields including manufacturing, aerospace, energy production, and construction. New York received three awards totaling $11,126,130: the New York State Department of Labor ($3,708,710), Per Scholas ($2,719,671), and Pursuit ($4,697,749).

Three days later, Governor Kathy Hochul announced the AI Jobs Institute in Long Island City, Queens, operated by Pursuit, backed by $10 million in total public investment, including $3.06 million from Empire State Development. It aims to train up to 40,000 New Yorkers by 2032, with partners including Google.org, OpenAI, Anthropic, Wells Fargo, and Adobe. Construction begins this winter, with full occupancy expected in mid-2027. Pursuit will also lead AI Prep for Jobs, which will train 420 low-income New Yorkers across the state. The release doesn’t describe Rochester or other upstate delivery for the institute itself.

• New York’s three Rapid Reskill awards total $11.1 million, the largest to Pursuit ($4.7 million).
• The grants set a six-month goal for placement in paid work, including apprenticeships and paid on-the-job training.
• The AI Jobs Institute targets 40,000 trainees in Queens by 2032; the statewide AI Prep for Jobs program covers 420.

The six-month placement goal signals where federal workforce policy is heading: faster movement into paid work, not open-ended training. For upstate workforce boards, the practical questions are where the state Labor Department’s $3.7 million will be used and whether AI Prep for Jobs seats reach Rochester. And 40,000 is a training target, not a placement count.

Impact: Emerging.

088. Employers Say Entry-Level Hiring Will Grow. The Data on Young Workers Hasn’t Turned.

A Korn Ferry report from its AMS subsidiary, released October 20, found that 81% of organizations plan to increase early-career hiring over the next two years, and 76% of talent acquisition leaders say entry-level jobs are not disappearing. Nearly three-quarters expect entry-level work to change rather than be eliminated, but roughly half don’t yet know what that change will look like. The research surveyed more than 500 HR and talent leaders and more than 5,000 job seekers, focused on ages 16 to 24, across 13 countries. AMS sells early-career recruiting services.

The day after, Stanford HAI described the latest readings from its Digital Economy Lab’s AI Economic Indicators dashboards: employment in AI-exposed jobs like software development and customer service “has dropped by more than 4% this year,” and for 22- to 25-year-olds in those jobs, the decline has recently sharpened to 4.4%.

• 81% of employers in Korn Ferry’s survey plan to increase early-career hiring over two years.
• Roughly half of those employers don’t yet know what the new entry-level job will look like.
• Stanford’s tracker shows employment for 22- to 25-year-olds in software and customer service roles falling 4.4%.

Entry-level hiring has been a frequent Career Intel theme; what’s new is this clean split between intentions and outcomes in the same week. Survey plans aren’t hires, and the most useful finding may be that half of employers haven’t designed the new entry-level role yet. For job seekers, it’s fair to ask in interviews what a first-year hire actually does now, and how they learn.

Impact: Emerging.

099. Closer to Home: MCC Wins $5.4 Million for Optics Training; West Main Grants Break Ground

Sen. Chuck Schumer and Rep. Joe Morelle announced on October 22 a three-year, $5,383,256 award to Monroe Community College through the Defense Department’s Manufacturing Engineering Education Program. It supports MCC’s Optical Systems Technology program: growing enrollment, adding training and apprenticeship opportunities, and strengthening the pipeline of technicians for optics, photonics, semiconductor, and defense employers. The award follows a $5.1 million grant in 2023, and the release cites more than 150 related companies in the Finger Lakes region.

On October 21, the state and OneROC held a ceremonial groundbreaking for the West Main Gateway Project in downtown Rochester, a $10 million grant fund split between the state’s Regional Revitalization Program and the Ralph C. Wilson Foundation. It offers building renovation grants of $25,000 to $250,000, small business equipment grants, and public space improvements, with 22 projects underway. A reminder from last week: New York’s unemployment system goes offline at 5 p.m. on Monday, November 2, for about 9 to 12 days, so claimants should file or certify before then.

• MCC’s $5,383,256 award funds more optics students plus new training and apprenticeship slots.
• The Finger Lakes region has more than 150 optics, photonics, and related companies, per the release.
• West Main Gateway grants cover up to 90% of eligible renovation costs for corridor businesses.

After a week of AI headlines, these are reminders that some of the most durable local pathways are technical and hands-on. Optics technician is a career that starts at a community college, not with a bachelor’s degree, which makes it relevant to the same workers Pew’s data flags in story 1. For career advisors, MCC’s expanded apprenticeship opportunities are worth tracking as they open.

Impact: Immediate.

Bottom line

For workers and job seekers, the lesson this week is to look past the average. Jobs exposed to AI are growing overall, but not for everyone, and women without degrees in administrative and customer service roles appear to be losing ground. Profitable companies are still cutting, low layoffs don’t mean easy hiring, and paychecks are trailing prices. The protection is the same as it has been: make your judgment and problem-solving visible, and look closely at pathways like MCC’s optics program that lead to technical work without a four-year degree.

For employers and workforce organizations, the signal is that totals aren’t enough. Track outcomes by gender and education, separate AI-driven cuts from strategy or tax-driven ones, and build transition capacity before the next downturn, which Goldman argues is when AI displacement will show up most. New federal and state training dollars are arriving in New York; the question for upstate is how much of them reaches Rochester.