5. Draft Analytical Report: Labor Market Baseline, AI Impact Assessment, Policy Recommendations — original pdf

Backup

Austin EPC | AI Policy Recommendation | DRAFT Austin Economic Prosperity Commission DRAFT ANALYTICAL REPORT Exploratory Phase Impact of Artificial Intelligence on Austin Residents: Labor Market Baseline, AI Impact Assessment, and Policy Recommendations July 2026 Commissioner Jake Randall (Policy Champion) Prepared for internal commissioner discussion consistent with EPC bylaws. Draft for Commissioner Review Page 1 Austin EPC | AI Policy Recommendation | DRAFT Table of Contents: 1. Executive Summary: ............................................................................................................... 3 2. Labor Market Baseline: ........................................................................................................... 4 2.1 Labor Market Conditions (MSA): ................................................................................................... 4 2.2 Industry Employment (MSA): ........................................................................................................ 4 2.3 Occupational Mix (MSA): .............................................................................................................. 5 2.4 City of Austin Resident Indicators: ................................................................................................ 7 2.5 Austin's Exposure Profile: ............................................................................................................. 7 3. AI Impact Assessment: ........................................................................................................... 8 3.1 Summary Assessment Table: ....................................................................................................... 8 3.2 Key Findings: ............................................................................................................................... 9 4. Policy Recommendations: .................................................................................................... 10 4.7 Sequencing: .............................................................................................................................. 12 Appendix A: Data Sources, Vintages, and Methods: ................................................................. 13 A.1 Sources and Vintages: ............................................................................................................... 13 A.2 Derivations: .............................................................................................................................. 13 A.3 Wage and Income Reconciliation: .............................................................................................. 13 A.4 Limitations and Open Questions: ............................................................................................... 13 Appendix B: Assessment Frameworks and Detailed Ratings: .................................................. 15 B.1 Frameworks: ............................................................................................................................. 15 B.2 Detailed Ratings: ....................................................................................................................... 15 References: .............................................................................................................................. 18 Draft for Commissioner Review Page 2 Austin EPC | AI Policy Recommendation | DRAFT 1. Executive Summary: This report delivers the labor market baseline, AI impact assessment, and policy recommendations directed by the March 2026 brief. Figures trace to named, dated public sources (Appendix A); City of Austin and Austin MSA data are labeled throughout; the occupational analysis covers major groups above 3% of MSA employment. Baseline: • ~1.41M nonfarm jobs (Apr 2026), +0.6% over the year; unemployment 3.4%. • Uneven growth: mining, logging, and construction +5.0%; information -2.6%; leisure and hospitality -2.1%; manufacturing -1.3%. • A large clerical base (office and administrative support, 12.4% of employment) sits alongside an outsized technical base (computer and mathematical, 6.3% vs 3.4% nationally; 28.2K software developers). • City residents: 60.7% hold a bachelor's degree or higher; median household income $90.4K; poverty 11.8%; 13.6% of under-65 residents uninsured. Highest-Priority Findings: 1. Office and administrative support: Austin's largest group, highly exposed in every framework, lowest tier of adaptive capacity. The MSA's largest downside concentration. 2. Computer and mathematical: split outcome. Augmentation and demand for experienced workers, but a ~20% national employment decline for software developers aged 22-25 since late 2022 (Stanford payroll data). 3. Business and financial operations (8.0% of employment): high exposure with high adaptive capacity. Expect role redesign and entry-path narrowing, not mass displacement. 4. Skilled trades tied to data center and electrical buildout are the clearest upside; local construction growth (+5.0%) is consistent with the channel. Top Three Recommendations: 1. AI Labor Market Dashboard with an early-career hiring module: entry-level hiring is where national evidence shows AI effects first. 2. Administrative and customer support transition compact: pair the largest high-exposure, low- capacity group with named destinations and wage-quality benchmarks. 3. Data center and electrical infrastructure skilled-trades pipeline: registered apprenticeships so residents capture infrastructure demand. Exposure is not displacement. Task change, hiring shifts, and entry-path narrowing are today's observable margins; broad job loss is not visible in top-line data, and the strongest research finds declines in exposed occupations statistically clear only from 2024 onward. The findings here are early signals to act on with monitoring and preparation, not settled outcomes. Draft for Commissioner Review Page 3 Austin EPC | AI Policy Recommendation | DRAFT 2. Labor Market Baseline: 2.1 Labor Market Conditions (MSA): Indicator Value (Apr 2026) Context Total Nonfarm Employment ~1.41M jobs Unemployment Rate Civilian Labor Force Average Weekly Wage 3.4% ~1.56M $1,878 +8.9K (+0.6%) over the year; growth slowed from +1.5% in Dec 2025 Down from 3.7% in Jan-Feb 2026; below the U.S. rate Roughly flat since Nov 2025 Q4 2025, QCEW; U.S. $1,569. Includes bonuses and irregular pay (Appendix A) Table 2.1. MSA labor market conditions. Source: BLS Economy at a Glance (LAUS/CES), April 2026 preliminary, not seasonally adjusted; extracted June 16, 2026. Key Insight: headline numbers show no broad stress. The growth slowdown and the sector split below are the signals to watch. 2.2 Industry Employment (MSA): Figure 2.1. MSA employment change by industry supersector. Source: BLS CES, April 2026 preliminary, via the Austin Area Economic Summary (June 3, 2026). Key Insight: the declines in information (-2.6%) and manufacturing (-1.3%) are plausibly cyclical and firm-specific; attributing them to AI would be premature. The fastest-growing supersector, mining, logging, and construction (+5.0%), is directionally consistent with the data center buildout channel, though CES cannot isolate data-center projects. Draft for Commissioner Review Page 4 2.3 Occupational Mix (MSA): Coverage: the 12 major occupational groups above 3.0% of MSA employment, ~84% of all jobs (shares per OEWS May 2024; derivation methods in Appendix A). Austin EPC | AI Policy Recommendation | DRAFT Major Occupational Group (SOC) Total, All Occupations Office and Administrative Support Management Employment (May 2024) ~1.27M 156.9K 129.0K Food Preparation and Serving Related 121.4K Sales and Related Business and Financial Operations Transportation and Material Moving Computer and Mathematical Educational Instruction and Library Healthcare Practitioners and Technical Construction and Extraction Installation, Maintenance, and Repair Production 111.3K 101.2K 84.8K 79.7K 67.0K 60.7K 56.9K 46.8K 43.0K Share of MSA Employment U.S. Share 100.0% 100.0% 12.4% 10.2% 9.6% 8.8% 8.0% 6.7% 6.3% 5.3% 4.8% 4.5% 3.7% 3.4% 11.8% 7.1% 8.8% 8.7% 6.7% 8.9% 3.4% 5.8% 6.2% 4.1% 3.9% 5.7% Share Ratio (ATX / U.S.) Annual Mean Wage 1.00 1.05 1.44 1.09 1.01 1.19 0.75 1.85 0.91 0.77 1.10 0.95 0.60 $71.4K $50.8K $144.2K $34.2K $56.7K $90.0K $44.4K $116.8K $63.5K $98.2K $55.6K $59.4K $47.2K Table 2.2. MSA occupational baseline, groups above 3% of employment. Source: OEWS May 2024 (BLS Austin regional release, June 17, 2025); counts, ratios, and annualized wages computed per Appendix A. BLS publishes a 1.88 location quotient for Computer and Mathematical (ratio shown: 1.85). Figure 2.2. Occupational concentration relative to the U.S. Source: computed from OEWS May 2024 shares. Draft for Commissioner Review Page 5 Austin EPC | AI Policy Recommendation | DRAFT Occupation (OEWS May 2025) Austin ($/Hour) U.S. ($/Hour) Computer and Information Systems Managers Computer Hardware Engineers Software Developers Postsecondary Education Administrators Public Relations Specialists Executive Secretaries and Exec. Admin. Assistants $92.53 $83.87 $69.05 $68.01 $36.52 $35.18 $92.39 $78.21 $71.20 $60.84 $40.44 $38.05 Table 2.3. Selected mean hourly wages, OEWS May 2025 (Austin Area Economic Summary, June 3, 2026). MSA all-occupations mean hourly wage: $35.85 vs $33.54 U.S. Key Insight: Austin software developer mean wages, above the national figure in May 2024 ($67.01/hour), fell below it in May 2025. One survey year cannot separate softening from composition change or sampling variation. Draft for Commissioner Review Page 6 2.4 City of Austin Resident Indicators: Geography: City of Austin except where noted. Household income aggregates all earners plus non-wage income and is not comparable to per-job wages (reconciliation in Appendix A). Austin EPC | AI Policy Recommendation | DRAFT Indicator Population Median Household Income Per Capita Income Poverty Rate High School Graduate or Higher, 25+ Bachelor's Degree or Higher, 25+ Median Gross Rent Under-65 Residents Without Health Insurance Value ~994K $90.4K $67.0K 11.8% 92.3% 60.7% $1,729 13.6% Source and Vintage ACS 2024 1-year ACS 2024 1-year; MSA $99.9K ACS 2024 1-year ACS 2024 1-year; MSA 9.2% ACS 2024 1-year ACS 2024 1-year; MSA 52.3%; U.S. 36.9% ACS 2020-2024 5-year ACS 2020-2024 5-year (Census QuickFacts convention) Median Value, Owner-Occupied Housing $571K ACS 2024 1-year Table 2.4. City of Austin resident indicators; 1-year and 5-year vintages labeled per row. 2.5 Austin's Exposure Profile: Key Insight: Austin is high-exposure, high-capacity in aggregate, with pockets where high exposure meets low capacity. Roughly 27% of MSA employment sits in the three most exposure-relevant groups (office and administrative support, business and financial operations, computer and mathematical); developers work here at 2.09x the national rate; management (1.44) and business and finance (1.19) are also overweight. National patterns of task change and entry-level hiring adjustment should surface here earlier than in manual-heavy metros. The same concentration is the labor pool that builds and governs AI systems, and it pays: the MSA mean hourly wage runs ~7% above the national mean (May 2025) and weekly wages ~20% above (Q4 2025). Education is a genuine cushion (60.7% bachelor's attainment vs 36.9% nationally), but only on average: 11.8% poverty, a 13.6% under-65 uninsured share, and $1,729 median rent mean a meaningful minority has thin buffers. Section 3 locates those pockets. Draft for Commissioner Review Page 7 3. AI Impact Assessment: Twelve major occupational groups were rated for exposure, likely effect, and confidence against seven published frameworks and 2025-26 empirical studies (frameworks and detailed ratings in Appendix B). Results first; the summary table follows. Austin EPC | AI Policy Recommendation | DRAFT 3.1 Summary Assessment Table: Group (MSA Share; Annual Mean Wage) Office and Administrative Support (12.4%; $50.8K) Management (10.2%; $144.2K) Food Preparation and Serving (9.6%; $34.2K) Sales and Related (8.8%; $56.7K) Business and Financial Operations (8.0%; $90.0K) Transportation and Material Moving (6.7%; $44.4K) Computer and Mathematical (6.3%; $116.8K) Educational Instruction and Library (5.3%; $63.5K) Healthcare Practitioners and Technical (4.8%; $98.2K) Construction and Extraction (4.5%; $55.6K) Installation, Maintenance, and Repair (3.7%; $59.4K) Production (3.4%; $47.2K) Exposure Likely Effect Confidence High Substitution risk dominant; hiring slowdown before layoffs High Medium Augmentation; slower middle-management growth Medium Low Minimal direct effect High Medium Substitution in remote-channel selling; augmentation in relationship sales Medium Augmentation for experienced staff; entry-path narrowing High / Medium Minimal near term; dispatch tasks partially exposed Medium Augmentation and demand growth for experienced workers; entry-level substitution pressure Medium Augmentation; minimal K-12 substitution High Medium High Low High Medium Augmentation plus demographic demand growth Medium Low Low Low Demand growth from data center and electrical buildout High / Medium Demand growth from data center operations High / Medium Near-term neutral; semiconductor cycle matters more Medium Table 3.1. Summary assessment, MSA major groups above 3% of employment. Shares and wages per Table 2.2 (OEWS May 2024). Framework justifications, adaptive capacity detail, and split confidence ratings in Appendix B, Table B.1. Draft for Commissioner Review Page 8 Austin EPC | AI Policy Recommendation | DRAFT 3.2 Key Findings: Largest Downside Concentrations (Weighted by Employment): 1. Office and Administrative Support (~156.9K Jobs): largest group, highly exposed in every framework, and the occupational core of the 6.1M U.S. workers who combine high exposure with bottom-quartile adaptive capacity. The near-term margin is hiring and attrition, not layoffs. 2. Customer Support and Structured Information-Handling Roles: customer service representatives are a Stanford exemplar occupation for early-career decline; remote-channel sales carries the highest exposure scores. 3. Early-Career Computer and Mathematical Workers (Group Total 79.7K): the group is a strength; the entry level is the documented pressure point (~20% national decline for developers aged 22-25 since late 2022, against stable experienced employment). Austin's developer concentration (2.09x national) makes this a first-order local concern. 4. Business and Financial Operations (~101.2K): high exposure, high capacity. Expect role redesign, rising productivity expectations, and slower junior hiring, concentrated in routine analysis and reporting. 5. Low-Wage Exposed Niches Within Office and Administrative Support: bookkeeping, data entry, and administrative assistant roles combine the highest task overlap with the thinnest buffers. Groups with Realistic Upside: • Computer and Mathematical, Experienced Tier: AI buildout, integration, security, and data work are demand channels, and complementarity rises with experience. The 79.7K-person base is Austin's main claim on this upside. • Construction and Skilled Trades (~103.7K Combined): buildout creates construction-phase demand (electricians, mechanical and cooling trades, controls, fiber) and permanent operations demand; the supersector's +5.0% year is directionally supportive. Resident capture magnitudes remain uncertain (Appendix A). • Implementation, Governance, and Training Roles: AI adoption adds integration, risk, data governance, and training capacity; Resolution 55 (Apr 24, 2025) implies public-sector demand. No occupational time series isolates these roles yet. Cross-Cutting Populations of Concern: Early-Career Workers and Recent Graduates: adjustment is happening first at the entry level, in automation-leaning occupations, through employment rather than wages. Austin's mix puts a disproportionate share of its young workforce in those occupations; if entry paths narrow while experienced employment holds, headline unemployment will understate the problem. Lower-Buffer Households in Exposed Occupations: a household near the office and administrative support annual mean wage of $50.8K faces $1,729 median rent, ~41% of that gross wage (a per-job wage against a household rent). For them, a moderate cut in hours is a solvency event. The policy- relevant population is the intersection of high exposure and thin buffers: clerical, customer support, and low-wage administrative workers. Draft for Commissioner Review Page 9 4. Policy Recommendations: Six recommendations prioritize long-term employment quality: wage levels, durable pathways, and resident capture of upside. Items requiring Council budget or ordinance authority are flagged. Austin EPC | AI Policy Recommendation | DRAFT 4.1 AI Labor Market Dashboard with an Early-Career Module: Element Detail Action Rationale Publish a quarterly dashboard of 8-10 AI-relevant indicators, with an early-career module: hiring, postings, and unemployment claims for workers under 30 in the five most exposed occupation families, benchmarked against experienced workers. Entry-level hiring is where payroll evidence shows AI effects first (-13% ages 22-25 in exposed occupations; ~-20% young software developers; Section 3.2). Without an age dimension, a dashboard misses the one documented margin. Implementers Economic Development Department (lead), Innovation Office, Workforce Solutions Capital Area, Texas Workforce Commission data, Austin Energy, a university research partner. Expected Effect Shared, non-alarmist evidence base; hiring inflections detected quarters before annual OEWS data. Resourcing Measurement Low. Existing public data plus staff time; standing publication requires Council direction or department initiative. Shipped within two quarters; indicator coverage; usage by Council offices and partners; detection lead time vs official data. 4.2 Administrative and Customer Support Transition Compact: Element Detail Action Rationale Implementers Expected Effect Resourcing Measurement A compact among Workforce Solutions Capital Area, Austin Community College (ACC), libraries, and 10-20 large employers of administrative and customer support staff: employers give advance notice of role redesign; training partners run short-cycle credentials toward named destinations (computer user support, medical administration, data and records governance, project coordination, trades pre-apprenticeship); placements target the worker's prior wage, floor- referenced to the group's $50.8K annual mean wage (Table 2.2). The region's largest group (~156.9K jobs) pairs the highest exposure with the lowest adaptive capacity, and rent absorbs ~41% of the group's annual mean wage (Section 3.2); transitions into lower-paying work fail the economic-security objective. Workforce Solutions Capital Area (lead), ACC, City workforce programs, Austin Public Library, employers, worker advocacy organizations. Higher transition rates and wage retention for the largest at-risk group; earlier employer signaling of role change. Medium. Mostly reallocated Workforce Innovation and Opportunity Act and ACC resources; new City funding requires Council action. Enrollment and completion by prior occupation; placement rate; wage replacement at 6 and 18 months (target 100%+); employer count. 4.3 Data Center and Electrical Infrastructure Skilled-Trades Pipeline: Element Detail Action Rationale A standing table (Austin Energy, permitting staff, contractors, apprenticeship sponsors, ACC) that publishes a rolling forecast of electrician, HVAC and cooling, controls, and fiber/network demand tied to announced projects; expands registered apprenticeship seats to match; and reserves pre- apprenticeship slots for workers transitioning under Recommendation 4.2. The two trades groups (~103.7K jobs combined) are low-exposure with a demand-growth channel; the construction supersector grew 5.0% over the year (Section 3.2). Annual mean wages of $55.6K and $59.4K make these durable, high-quality non-degree pathways. Draft for Commissioner Review Page 10 Austin EPC | AI Policy Recommendation | DRAFT Element Detail Implementers Expected Effect Resourcing Measurement Austin Energy and Economic Development Department (co-leads), ACC, apprenticeship sponsors, contractor and trades partners, regional workforce entities. Higher resident capture of infrastructure demand; a high-wage destination inside the transition compact. Medium. Apprenticeship expansion draws state and federal funds; City incentives require Council action. Seats and completions by trade; resident share of completers; entrants from exposed occupations; journeyworker wages; forecast accuracy vs permitted projects. 4.4 Early-Career Pathways Initiative: Element Detail Action Rationale Two parts: (a) paid, structured entry-level cohorts inside City departments (analyst, technology, administrative modernization) that use AI tools as training instruments rather than entry-role replacements; (b) an employer compact reporting entry-level hiring intentions annually and piloting apprenticeship-style entry roles in software, data, and financial operations, with UT Austin and ACC aligning capstone and co-op programs. The documented pressure point is entry-level employment in exposed occupations (Section 3.2). If firms stop training juniors, the experienced workforce behind Austin's AI upside erodes over a decade; this targets pathway durability. Implementers City Human Resources and Innovation Office (cohorts; requires Council budget action); Economic Development Department and Austin Chamber (compact); UT Austin, ACC, Workforce Solutions Capital Area. Expected Effect Preserved career ladders in Austin's central occupations; a visible public-sector standard for AI-era entry roles. Resourcing Medium for cohorts (Council action required); low for the compact. Measurement Cohort seats and conversion-to-permanent rates; compact employer count and reported hiring trends; regional under-30 hiring vs national benchmarks (dashboard, 4.1). 4.5 City AI Adoption Workforce Review with Redeployment Standards: Element Detail Action Rationale Adopt a workforce review for significant City AI procurements with two standards: a redeployment- first rule (retraining and internal mobility plans precede any headcount reduction) and a short annual public report on deployments, tasks changed, and workforce outcomes, per Resolution 55 (Apr 24, 2025). The City is itself a large administrative employer with the exposure profile flagged in Section 3.2. Documented internal practice makes City guidance to private employers credible and creates the governance roles noted among upside channels. Implementers Innovation Office (lead), Procurement, Human Resources, Law, departmental owners. Formal procurement thresholds require Council action; an administrative pilot does not. Expected Effect Disciplined adoption; preserved employment quality for City staff; a copyable local model. Resourcing Low to medium (staff time; no new systems). Measurement Share of qualifying procurements reviewed; staff retrained or redeployed vs separated; report published on schedule. 4.6 AI-Complementary Skills Inside Existing Credentials: Element Action Detail Embed applied AI-tool fluency inside existing credentials (administrative technology, bookkeeping, customer operations, computer support, trades estimating and controls) rather than standalone Draft for Commissioner Review Page 11 Austin EPC | AI Policy Recommendation | DRAFT Element Detail Rationale generic AI courses; pair with library-based tool access and coaching for residents without employer- provided access. Over half of observed AI work interactions are augmentation, and complementarity rises with experience (Appendix B). Embedding fluency in occupational credentials reaches the exposed workers identified in Section 3.2; open access counters the risk that fluency concentrates among the already advantaged. Implementers ACC (lead), Workforce Solutions Capital Area, Austin Public Library, employer partners for curriculum grounding. Expected Effect Within-occupation resilience; stronger candidates for the destinations in 4.2 and 4.3. Resourcing Low to medium (curriculum revision within existing programs; library programming). Measurement Credentials with embedded components; enrollees from priority occupations; completer employment and wage outcomes vs prior cohorts. 4.7 Sequencing: The dashboard (4.1) costs least, needs no new authority, and calibrates the rest; 4.2 and 4.3 interlock as the core worker-facing pair; 4.5 can begin administratively at any time; 4.4 and 4.6 are the long- horizon pathway investments to scope during the dashboard's first two quarters. If Council supports only three actions: 4.1, 4.2, 4.3. Draft for Commissioner Review Page 12 Austin EPC | AI Policy Recommendation | DRAFT Appendix A: Data Sources, Vintages, and Methods: A.1 Sources and Vintages: • Occupational Data: BLS published May 2025 OEWS estimates on May 15, 2026, but the complete Austin MSA major-group table is available only through an interactive application that blocks automated retrieval; the Austin regional release remains May 2024 (June 17, 2025). This report uses May 2025 where verifiable in static publications (Austin Area Economic Summary, June 3, 2026) and May 2024 for the full major-group table; vintages are never mixed within a row. • Labor Market Conditions: CES and LAUS, April 2026 preliminary, not seasonally adjusted, extracted June 16, 2026. • Resident Indicators: ACS 2024 1-year estimates where available; ACS 2020-2024 5-year for median gross rent and the under-65 uninsured rate. Vintages labeled per row. • Geography: labor data cover the five-county Austin MSA; resident indicators cover the City of Austin. MSA occupational shares are an imperfect but best-available proxy for city residents; every table and chart is labeled City or MSA. A.2 Derivations: • Employment Counts: published share x implied OEWS total of ~1.27M jobs (from 79.7K computer and mathematical jobs = 6.3%), rounded. • Share Ratio: Austin share / U.S. share, a proxy for the location quotient (published LQ for computer and mathematical: 1.88 vs ratio 1.85). • Annual Wages: hourly mean x 2,080 hours (BLS convention). Group medians appear only in files not retrievable at execution time and are omitted. A.3 Wage and Income Reconciliation: Four correct figures could be misread as conflicting; each measures a different unit, pay concept, or geography. Figure Value What It Measures OEWS Annual Mean Wage, All Occupations (MSA) $71.4K (May 2024); $74.6K (May 2025) Wage per job, annualized at 2,080 hours; excludes irregular pay such as bonuses and stock. Per worker, not per household. QCEW Average Weekly Wage (MSA Counties) $1,878 (Q4 2025) Total covered payroll over covered jobs; includes bonuses and exercised stock; Q4 is seasonally elevated. The gap with the OEWS mean is definitional. Median Household Income, City of Austin $90.4K (ACS 2024 1- year) All earners in a household plus non-wage income; a household median can exceed a per-job mean. Median Household Income, Austin MSA $99.9K (ACS 2024 1- year) Same concept, wider geography; suburban counties have higher medians. Table A.1. Wage and income figures reconciled. Poverty (11.8%) is ACS 2024 1-year; rent ($1,729) and the under-65 uninsured rate (13.6%) are ACS 2020-2024 5-year averages, which lag current conditions. A.4 Limitations and Open Questions: • Vintage: baseline shares are May 2024; retrievable May 2025 toplines suggest no material structural change. Refresh when the regional release publishes (expected mid-2026). • Derived Figures: computed counts, ratios, and annualized wages carry rounding error of a few hundred jobs per group; adequate for prioritization, not budgeting. Draft for Commissioner Review Page 13 Austin EPC | AI Policy Recommendation | DRAFT • Attribution: no local series isolates AI from interest rates, the technology cycle, or restructuring; the strongest research finds clean attribution only from 2024 on. Findings are monitoring priorities, not causal claims. • Framework Limits: exposure indices track current model capability; usage data reflect one vendor; payroll data one provider. Convergence across independent methods mitigates but does not eliminate this. • Unverified Carry-Over: the 124 MW Austin Energy data center figure is carried from the March 2026 brief; re-confirm during scoping of Recommendation 4.3. No other figure is carried without verification. • Open Questions: whether Texas Workforce Commission wage records can support an age- disaggregated local hiring series; what share of data center jobs go to Austin residents; how large employers are changing entry-level hiring standards. Draft for Commissioner Review Page 14 Austin EPC | AI Policy Recommendation | DRAFT Appendix B: Assessment Frameworks and Detailed Ratings: B.1 Frameworks: • AIOE (Felten, Raj, and Seamans, 2021): occupational exposure index; highest in cognitive, analytical, communication-heavy work, lowest in manual work. • GPTs are GPTs (Eloundou et al., 2023; Science 2024): ~80% of U.S. workers have 10%+ of tasks exposed to large language models, ~19% have 50%+; exposure rises with wage and education. • Anthropic Economic Index (Feb 2025 through Jun 2026): measured AI usage concentrates in software and technical writing; augmentation exceeds automation (57% initial; 52% Nov 2025); computer and mathematical tasks are ~1/3 of consumer conversations, ~1/2 of business API traffic. • Anthropic, Labor Market Impacts of AI (2026): usage-based exposure measure; used as a cross- check. • Brookings / Centre for the Governance of AI Adaptive Capacity (Manning and Aguirre, 2026): of 37.1M highly exposed U.S. workers, ~70% have above-median adaptive capacity; 6.1M are highly exposed with bottom-quartile capacity, concentrated in clerical work. • Canaries in the Coal Mine (Brynjolfsson, Chandar, and Chen; Stanford, Aug 2025 to Feb 2026): ~13% relative employment decline for ages 22-25 in the most exposed occupations since late 2022 (~20% for young software developers), concentrated in automation-leaning occupations; the Feb 2026 update finds declines cleanly attributable only from 2024 onward. • Brookings Geography of Generative AI (2023-24): exposure concentrates in high-wage, high- education metros, Austin included. B.2 Detailed Ratings: Group (MSA Share; Annual Mean Wage) Exposure Framework Justification Direction of Likely Effect Adaptive Capacity Confidence Office and Administrative Support (12.4%; $50.8K) High Clerical and routine information tasks rate high in AIOE, GPTs are GPTs, and AEI usage Substitution risk dominant; some augmentation; hiring slowdown more likely than layoffs near term Low: clerical roles anchor the bottom quartile of the adaptive capacity index; moderate wages High: all frameworks agree; Canaries finds automation- type declines here Management (10.2%; $144.2K) Medium Analytical and writing tasks exposed; interpersonal, supervisory, and judgment tasks less so Augmentation dominant; possible span-of-control widening that slows middle-management growth High: high wages, dense local market, transferable skills Medium: heterogeneous task mix Food Preparation and Serving Related (9.6%; $34.2K) Low Physical, in-person tasks rate low in all task-based frameworks Minimal direct effect; indirect exposure via local demand Low buffers (lowest group wage in the MSA), but low exposure limits AI-specific risk High Sales and Related (8.8%; $56.7K) Medium Bifurcated: in-person retail low; phone, chat, and inside sales among the highest AIOE scores Substitution in scripted remote-channel selling; augmentation for relationship sales Mixed: retail has thin buffers; B2B sales has high wages and transferable skills Medium Business and Financial Operations (8.0%; $90.0K) Transportation and Material Moving High Low Draft for Commissioner Review Analysis, compliance, and reporting rate high in AIOE and GPTs are GPTs; heavy AEI usage Augmentation dominant for experienced staff; substitution risk in routine analysis; entry-path narrowing per Canaries High: financial occupations score near the top of the capacity index despite high exposure High for exposure; Medium for net employment effect Physical tasks rate low in generative-AI Minimal near term; dispatch and logistics Low to medium: modest wages and buffers; Medium: technology Page 15 Group (MSA Share; Annual Mean Wage) (6.7%; $44.4K) Computer and Mathematical (6.3%; $116.8K) High Educational Instruction and Library (5.3%; $63.5K) Medium Exposure Framework Justification Direction of Likely Effect Adaptive Capacity Confidence Austin EPC | AI Policy Recommendation | DRAFT frameworks; autonomous vehicles are a separate channel outside scope Highest measured usage of any group (AEI); high in AIOE and GPTs are GPTs Teaching scores high in AIOE, but delivery is in- person and publicly funded; preparation and grading exposed coordination partially exposed skills transfer within logistics channel uncertainty All three directions coexist: augmentation of experienced developers; demand growth in AI buildout roles; entry-level substitution pressure (~20% decline, developers 22-25, national) Augmentation dominant; minimal K-12 substitution; tutoring-adjacent roles more contestable High overall (top wages, dense market, transferable skills); early-career workers face the narrowest entry paths High: usage, exposure, and payroll evidence align Medium: stable public demand, credentialed skills, moderate wages Medium Healthcare Practitioners and Technical (4.8%; $98.2K) Construction and Extraction (4.5%; $55.6K) Installation, Maintenance, and Repair (3.7%; $59.4K) Production (3.4%; $47.2K) Medium Documentation, coding, and decision support exposed; clinical care, licensure, and liability limit substitution Augmentation plus demand growth from aging; below-average share (ratio 0.77) implies room to grow High: licensure, high wages, shortage conditions Medium Low Low Low Physical site work rates low everywhere; estimating and scheduling partially exposed Demand growth: data center, electrical, and cooling buildout; supersector +5.0% over the year (Apr 2026) Medium: apprenticeship pathways; mid-range wages; the main risk is the cycle, not AI High for low exposure; Medium for demand magnitude Hands-on diagnostic and repair work rates low; AI diagnostic tools augment rather than replace Physical production rates low in generative- AI frameworks; industrial robotics is an older, separate channel Demand growth: data center operations need electricians, HVAC and cooling, controls, and network trades on an ongoing basis Near-term neutral from generative AI; local semiconductor investment matters more Medium: credentialed trades transfer well; above-median wages for non-degree pathways Low to medium: share already well below national (0.60); plant- specific skills transfer imperfectly High for low exposure; Medium for demand Medium Table B.1. Detailed assessment ratings. Shares and wages per Table 2.2 (OEWS May 2024). Draft for Commissioner Review Page 16 References: Austin EPC | AI Policy Recommendation | DRAFT 1. U.S. Bureau of Labor Statistics, Occupational Employment and Wages in Austin-Round Rock-San Marcos, May 2024. News release 25-922-DAL, June 17, 2025. https://www.bls.gov/regions/southwest/news- release/occupationalemploymentandwages_austin.htm 2. U.S. Bureau of Labor Statistics, OEWS May 2025 estimates, published May 15, 2026. https://www.bls.gov/oes/tables.htm 3. U.S. Bureau of Labor Statistics, Austin Area Economic Summary, updated June 3, 2026 (OEWS May 2025 selected wages; CES Apr 2026; LAUS Apr 2026; QCEW Q4 2025). https://www.bls.gov/regions/southwest/summary/blssummary_austin.pdf 4. U.S. Bureau of Labor Statistics, Economy at a Glance: Austin-Round Rock-San Marcos, TX, data through April 2026 (preliminary), extracted June 16, 2026. https://www.bls.gov/eag/eag.tx_austin_msa.htm 5. U.S. Census Bureau, American Community Survey 2024 1-year estimates, City of Austin and Austin MSA (released Sept 2025), via Census Reporter. http://censusreporter.org/profiles/16000US4805000-austin-tx/ 6. U.S. Census Bureau, ACS 2020-2024 5-year estimates (released Dec 2025), City of Austin. https://www.census.gov/quickfacts/fact/table/austincitytexas 7. Felten, E., Raj, M., and Seamans, R., Occupational, Industry, and Geographic Exposure to Artificial Intelligence. Strategic Management Journal, 2021 (AIOE index). 8. Eloundou, T., Manning, S., Mishkin, P., and Rock, D., GPTs are GPTs. 2023; updated in Science, 2024. 9. Anthropic, The Anthropic Economic Index, first report, Feb 2025. https://www.anthropic.com/news/the-anthropic- economic-index 10. Anthropic, AEI report: Economic Primitives, Jan 2026 (Nov 2025 data). https://www.anthropic.com/research/anthropic-economic-index-january-2026-report 11. Anthropic, AEI reports, Mar 2026 and Jun 2026. https://www.anthropic.com/research/economic-index-march- 2026-report and https://www.anthropic.com/research/economic-index-june-2026-report 12. Anthropic, Labor Market Impacts of AI, 2026. https://www.anthropic.com/research/labor-market-impacts 13. Brynjolfsson, E., Chandar, B., and Chen, R., Canaries in the Coal Mine? Stanford Digital Economy Lab, Aug 2025; Nov 2025 revision; Feb 2026 update. https://digitaleconomy.stanford.edu/publications/canaries-in-the-coal-mine/ 14. Manning, S., and Aguirre, T., How Adaptable Are American Workers to AI-Induced Job Displacement? NBER chapter; Brookings summary, Jan-Feb 2026. https://www.brookings.edu/articles/measuring-us-workers-capacity- to-adapt-to-ai-driven-job-displacement/ 15. Brookings Institution, The Geography of Generative AI's Workforce Impacts, 2023-24. 16. City of Austin, Council Resolution adopted Apr 24, 2025, agenda item 55, as cited in the March 2026 EPC brief. 17. Austin Energy, Resource, Generation and Climate Protection Plan to 2035 (124 MW service-area data center demand), as cited in the March 2026 EPC brief; not independently re-verified. 18. Austin Economic Prosperity Commission, Impact of Artificial Intelligence on Austin Residents, draft policy recommendation, ideation phase, March 2026. Draft for Commissioner Review Page 17