Tracing the impact of AI on the US labor market

Tracking the effects of AI on theUS labor market (May 2026) Matthew Luzzetti, Amy Yang, Justin Weidner,Sourav Dasgupta, Suvir RanjanDeutsche Bank Center, 1 Columbus CircleNew York, New York 10019Tel: 212 250 6161 1. Firm AI adoption Most sectors have well below 50% AI adoption, but adoption isexpected to rise Larger firms have much higher rates of AI adoption; generallytrue across sectors as well Significant variation in cross-state AI adoption, with highadoption states nearly 4x low adoption; national avg. near 20% Firms using AI are mostly using it to “supplement or enhance”employees Significant share of firms have not yet changed employmentlevels due to AI In the last six months, did this business use AI to do any of thefollowing? Firms with higher AI adoption are more likely to use AI toperform an existing task than introduce a new task 2. Developments in job Total job postings been broadly stable since late 2025 Job openings in some AI-sensitive occupations have fallenfaster than overall in past few years but have stabilized Some divergence in job opening trends for occupations thathave been a market worry recently No clear relationship between change in job postings overthe next year and substitutability with AI Share of job postings containing AI terms is still low butrising sharply in recent months Over 2025, AI exposed jobs saw the largest share andchange in job postings mentioning AI Jobs with AI mentions in knowledge work sectors Source:Indeed, Deutsche Bank Share of job postings referencing AI is rising across countries 3. Layoffs Limited evidence of a change in the low-firing environmentacross key layoffs indicators WARN indicators show no broad-based impact of AI aslayoffs remain low Limited relationship between AI adoption rate and year-over- Initial jobless claims also remain low No relationship between AI adoption rates across states andchange in initial jobless claims Continuing claims have moderated and are near the lows ofthe past three years, though they are higher than in ‘19 No relationship between state-level AI adoption rate andyearly change in continuing claims Jobless claims data versus the past few years in NSA terms–initial claims low, continuing in the middle WARN indicators are somewhat higher in states where firmsare more likely to use AI to replace tasks No relationship between initial claims and firms using AI toreplace tasks Limited relationship between continuing claims and statesusing AI to replace tasks 4. Wages Wage growth remains the highest for bachelor’s degree orhigher Limited relationship between wage growth and AI adoptionaccording to average hourly earnings Sectors with higher AI adoption have shown more significantdeceleration in wage growth on average Younger workers (16-24 ages) have shown some uptrend inwage growth in recent months Wage premium persists for college degrees 5. Labor developments for More noticeable rise in unemployment rates for younger agegroups, though some improvement since last year But employment-to-population ratios for younger individualsare not far from 2019 levels Unemployment rate for young workers (16-24) w/ bachelor’sdegree is elevated, while other education groups are not Unemployment rate for younger workers with bachelor’sdegree looks elevated relative to 2000-2019 history NY Fed finds that unemployment rates are up for recentcollege graduates And underemployment rate near the highest in a few years Appendix 1Important Disclosures *PricesarecurrentasoftheendoftheprevioustradingsessionunlessotherwiseindicatedandaresourcedfromlocalexchangesviaReuters,Bloombergandothervendors.OtherinformationissourcedfromDeutscheBank,subjectcompanies,andothersources.ForfurtherinformationregardingdisclosuresrelevanttoDeutscheBankResearch,pleasevisitourglobaldisclosurelook-uppageonourwebsiteathttps://research.db.com/Research/Disclosures/FICCDisclosures. Analyst Certification Theviewsexpressedinthisreportaccuratelyreflectthepersonalviewsoftheundersignedleadanalyst(s).Inaddition,theundersignedleadanalyst(s)hasnotandwillnotreceiveanycompensationforprovidingaspecificrecommendationorviewinthisreport.MatthewLuzzetti AdditionalInformation TheinformationandopinionsinthisreportwerepreparedbyDeutscheBankAGoroneofitsaffiliates(collectively'DeutscheBank').Thoughtheinformationhereinisbelievedtobereliableandhasbeenobtainedfrompublicsourcesbelievedtobereliable,DeutscheBankmakesnorepresentationastoitsaccuracyorcompleteness.Hyperlinkstothird-partywebsitesinthisreportareprovidedforreaderconvenienceonly.DeutscheBankneitherendorsesthecontentnorisresponsiblefortheaccuracyorsecuritycontrolsofthosewebsites. IfyouusetheservicesofDeutscheBankinconnectionwithapurchaseorsaleofasecuritythatisdiscussedinthisreport,orisincludedordiscussedinanothercommunication(oralorwritten)fromaDeutscheBankanalyst,DeutscheBankmayactasprincipalforitsownaccountorasagentforanotherperson. DeutscheBankmayconsiderthisreportindecidingtotradeasprincip