Item 3- AI Labor Policy- Exploratory Phase — original pdf

Backup

Austin EPC | AI Policy Recommendation | DRAFT Austin Economic Prosperity Commission DRAFT POLICY RECOMMENDATION Ideation Phase Impact of Artificial Intelligence on Austin Residents: Labor Market and Economic Security March 2026 Commissioner Jake Randall (Policy Champion) Commissioner Tomaszewski Commissioner Joshi Commissioner Rashed Prepared for internal commissioner discussion. This document does not represent a formal position of the Economic Prosperity Commission. Submitted for Commissioner review and working-group discussion consistent with EPC bylaws. Draft for Commissioner Review Page 1 Austin EPC | AI Policy Recommendation | DRAFT Table of Contents 1. Executive Summary 2. Impact Snapshot (Preliminary) 3. Relevant Context 4. Glossary of Key Terms 5. Supporting Detail and Analysis 5.1 Problem Statement 5.2 Three Analytical Components Within the Ideation Phase 5.3 Preliminary Impact Hypothesis, Baseline Findings, and Measurement Plan 5.4 Feasibility Constraints 5.5 Policy and Program Options 5.6 Key Risks and Mitigations 6. Stakeholder Signals and Map (External) 7. Recommended Next Steps (60-Day Plan) 7.1 Days 1–15: Baseline Assembly 7.2 Days 15–40: Impact Estimation and Stakeholder Input 7.3 Days 40–60: Policy Design and Synthesis 7.4 Evidence Status: Verified, Preliminary, or Removed 8. References Draft for Commissioner Review Page 2 Austin EPC | AI Policy Recommendation | DRAFT 1. Executive Summary Austin should approach AI as a labor-market issue first, not only as a technology issue. The key policy question is straightforward: how is AI likely to affect the labor market and economic security of Austin residents, and how should the City prepare? Within the Commission’s construction and job creation mandate, this brief focuses primarily on job creation, job retention, and resident economic security, while treating construction as a secondary but relevant channel where AI-related infrastructure may increase local skilled-trades demand. Austin enters the AI paradigm shift with both strengths and vulnerabilities. The Austin metro has a large professional and technical workforce, low unemployment, and a durable base in computer and mathematical work. The city also has a highly educated population and high incomes. At the same time, Austin has many residents in office support, sales, food service, customer-facing, and other occupations where task change can create uneven effects across households. The current evidence does not justify alarmist claims about large, immediate job loss. Published research is more cautious: AI is likely to change tasks first, especially in cognitive and clerical work, before broad labor-market effects become visible. Early evidence also suggests that exposure is different from displacement. Some workers in highly exposed occupations may adapt successfully. Others may not, especially workers with weaker savings, narrower skill transfer options, or fewer retraining pathways. Austin may also benefit from AI. AI can support local job creation and economic activity through technical work, implementation and training work, policy and oversight functions, and infrastructure-related construction. Austin’s large computer and mathematical workforce and the region’s ongoing digital infrastructure make that upside plausible. The Commission should therefore treat AI as both a risk- management issue and a readiness issue. This brief recommends a three-phase work program: 1. Baseline Assessment. Build a clear high-level picture of Austin’s workforce and resident economic conditions. 2. Impact Estimation. Identify which groups of residents may be most impacted by AI-driven task change, and where the city may also see new demand. 3. Policy Proposals. Develop practical city actions that improve preparedness without overstating what is known. Target Outcome: A focused, evidence-based recommendation to Council on how the City of Austin can monitor, prepare for, and respond to AI-related labor-market change affecting residents. Draft for Commissioner Review Page 3 Austin EPC | AI Policy Recommendation | DRAFT 2. Impact Snapshot (Preliminary) Focus Area Estimate Metric Impact Score Confidence Skilled trade demand; infrastructure-related construction activity 2 Low Occupational exposure; hiring; wages; claims; training access 3 Medium Construction Job Creation AI-related digital infrastructure may add targeted demand for electricians, HVAC and cooling, controls, networking, and maintenance trades. Near-term local construction effects appear plausible but modest at this stage. AI is more likely to change tasks first in selected cognitive and clerical occupations, while also supporting some job growth in technical, implementation, governance, and training roles. Net local effects remain uncertain and should be monitored. Note: This ideation-phase snapshot follows the Commission’s Construction and Job Creation impact pillars. The brief’s focus is the Job Creation pillar, with construction treated as a secondary impact channel where supported by evidence. 3. Relevant Context EPC Bylaws direct the Commission to advise Council on construction and job creation. AI falls within that mandate, but the labor question should be framed carefully. • The City already recognizes AI as a workforce issue. In April 2025, City Council adopted Resolution 55 on artificial intelligence guidelines, oversight, workforce considerations, environmental impacts, digital equity, and public engagement. • Austin’s labor market is large and still relatively tight. The Austin metro ended 2025 with about 1.389 million nonfarm jobs and a 3.2% unemployment rate. • Austin has a strong white-collar and technical base. In the Austin metro, office and administrative support is the largest occupational group at 12.4% of employment. Management is 10.2%. Food preparation and serving is 9.6%. Sales are 8.8%. Business and financial operations are 8.0%. Computer and mathematical occupations are 6.3%, well above the national share. Draft for Commissioner Review Page 4 Austin EPC | AI Policy Recommendation | DRAFT • Austin residents are highly educated, but not uniformly secure. In the city, 59.6% of adults aged twenty-five and older hold a bachelor’s degree or higher. Median household income is $93,658. At the same time, 12.0% of residents live in poverty and 13.6% of residents under age 65 lack health insurance. • AI exposure is different from job loss. The published literature consistently distinguishes task exposure from actual displacement. That distinction is important for local policy. Exposure can mean substitution, augmentation, work redesign, or productivity gain. • Austin also has job-creation pathways. Austin Energy’s planning materials indicate that current data-center demand in its service area is meaningful already and could grow further over time. That matters for local construction, electrical, cooling, networking, maintenance, and related support work. Given all of this, the central policy question is which Austin residents are most likely to face downside risk, which are most likely to capture upside, and what the City can do now before stronger labor effects become visible. Draft for Commissioner Review Page 5 Austin EPC | AI Policy Recommendation | DRAFT 4. Glossary of Key Terms Term Definition AI (Artificial Intelligence) Software systems that can perform or help perform tasks such as writing, coding, analysis, classification, prediction, and content generation. AI Exposure Task Change Augmentation Automation Adaptive Capacity Economic Security Data Center The degree to which the tasks in an occupation overlap with what current AI systems can do well. Exposure is different from job loss. A change in how work is conducted. Some tasks may be automated, sped up, reassigned, or improved, while the job itself still exists. Use of AI to assist a worker rather than replace the worker’s role. Use of AI to complete a task with limited human input. A worker’s practical ability to adjust if job tasks change or a job is lost. May depend on savings, transferable skills, age, local job options, and access to retraining. A resident’s practical ability to maintain earnings, meet basic costs, and absorb disruption without severe hardship. A facility that houses computing and networking infrastructure. In this brief, data centers matter mainly because they can affect local construction, utility demand, and some skilled-trades employment. Draft for Commissioner Review Page 6 Austin EPC | AI Policy Recommendation | DRAFT 5. Supporting Detail and Analysis 5.1 Problem Statement Austin does not yet have a shared, City-relevant framework for understanding how AI may affect residents through the labor market. Four gaps matter most: • No common baseline. There is no concise working baseline that brings together resident-level data for the City of Austin and labor-market data for the Austin metro. • No Austin-specific exposure view. The city does not yet have a simple, defensible estimate of which local occupations and resident groups may be most exposed to AI-driven task change. • No standing monitoring system. The city does not have a regular way to watch for early signals such as hiring slowdown, occupational shifts, claims patterns, or wage pressure in exposed categories. • No practical city posture. The city has begun AI governance work, but it has not yet translated that work into a narrow labor and economic-security preparedness agenda for residents. This brief is an initial step toward closing those gaps. The document is scoped to what currently available evidence supports and does not attempt to reach beyond that. 5.2 Three Analytical Components Within the Ideation Phase This revised brief is organized around three analytical components within the Commission’s Ideation phase. Under the Commission’s formal workflow, initiatives then move through Exploratory, Drafting, and Finalization. Phase 1: Baseline Assessment Establish a high-level picture of Austin’s current workforce and resident economic conditions. This phase should answer basic questions: What kinds of work are common in Austin? How educated is the resident base? Where are wages high or low? Which parts of the population may have less economic buffer if work changes quickly? This phase should use City of Austin resident data where possible, and Austin metro labor data where that is the most authoritative source. Phase 2: Impact Estimation Estimate, at a high level, which portions of Austin’s population and workforce may be most affected by AI. This phase should not attempt to forecast exact job losses, but instead identify areas of task exposure, low adaptive capacity, hiring slowdown, uneven upside versus downside, and realistic job-creation channels. This is where Austin’s occupational mix should be translated through published AI-exposure research. Draft for Commissioner Review Page 7 Austin EPC | AI Policy Recommendation | DRAFT Phase 3: Policy Proposals Develop practical policy proposals the City could consider. This phase should focus on actions the city can influence monitoring, training alignment, workforce transition support, city procurement and deployment standards, and infrastructure coordination where local job demand may grow. 5.3 Preliminary Impact Hypothesis, Baseline Findings, and Measurement Plan Hypothesized Mechanism of Action The working hypothesis is that AI may affect Austin residents through two main channels. 1. First, AI can change or automate selected cognitive and clerical tasks, which may affect hiring, wages, and job stability in exposed occupations. 2. Second, AI adoption and related infrastructure buildout can create or expand demand in selected technical, implementation, and skilled-trades roles. These channels map to the Commission’s Job Creation and Construction pillars. A. Baseline Findings The following baseline is intentionally high-level. Topic High-Level Finding Why It Matters Workforce size and labor-market conditions The Austin metro labor market remains large and tight, with about 1.4 million nonfarm jobs and 3.2% unemployment at the end of 2025. Broad labor-market stress is not yet visible in top-line numbers. Early AI effects may show up first in hiring or task redesign rather than headline unemployment. Major industry groups Occupational mix Professional and business services is one of the region’s largest industry groups, at about 276,100 jobs. Trade, transportation, and utilities are about 219,600. Education and health services is about 169,300. Leisure and hospitality levels are about 149,200. Construction is about 90,200. Information is smaller in count, at about 47,800, but still economically important. Office and administrative support accounts for 12.4% of metro employment. Management is 10.2%. Food preparation and serving is 9.6%. Sales are 8.8%. Business and financial operations are 8.0%. Computer and mathematical occupations are 6.3%. Austin has both high-exposure office- based industries and lower-exposure in- person sectors. This creates a mixed local pattern rather than a single citywide effect. This mix matters because generative AI appears most capable in cognitive and clerical tasks, while many in-person manual roles remain less exposed for now. Technical base Austin has 79,720 computer and mathematical jobs. Software developers Austin is well positioned to capture some AI-related technical upside, but that same Draft for Commissioner Review Page 8 Austin EPC | AI Policy Recommendation | DRAFT Education alone account for 28,210 jobs. Computer user support specialists account for 8,520, and computer systems analysts account for 8,310. In the city, 92.3% of adults age 25 and older have completed high school, and 59.6% hold a bachelor’s degree or higher. concentration also means more residents work in occupations where AI tools are already being used heavily. A highly educated labor force may adapt better on average, but it also means Austin has many residents in the kinds of knowledge work that AI can affect. Income and household resilience City median household income is $93,658, but poverty remains 12.0%. Median gross rent is $1,729, and 13.6% of residents under age 65 lack health insurance. Austin’s top-line prosperity does not mean all households can absorb a period of lower hours, lower earnings, or longer job search. Infrastructure and buildout Austin Energy planning materials estimate current service-area data-center demand at 124 MW and suggest additional demand could rise over time. That does not prove a local jobs boom, but it does support careful attention to skilled- trades demand linked to electrical, cooling, and infrastructure work. B. Preliminary Impact Estimation A careful Austin-specific reading of the literature suggests five early observations. 1. Task change is more likely than immediate broad job loss. Published AI exposure studies are best read as task-based warnings, not direct forecasts of unemployment. 2. Austin’s professional workforce raises local exposure. High shares of office, analysis, finance, software, and administrative work suggest Austin may see earlier workplace change than cities with more manual occupational mixes. 3. Exposure is likely to be uneven among residents. Workers in high-exposure occupations who also have weak financial buffers, narrow skills, limited credentials, or low access to retraining may face the greatest downside. 4. Young workers deserve special attention. Early evidence from recent research suggests that in some exposed occupations, slower hiring may appear before unemployment does. 5. Job creation should be tracked alongside disruption. Austin’s technical base and regional infrastructure pipeline mean the city may also capture gains in software, systems work, cybersecurity, training, AI operations support, and selected skilled trades. C. Resident Groups and Work Categories to Watch This section identifies where the city should pay closer attention. Group or Category Why It May Matter Office and administrative workers This is Austin’s largest occupational group. The published literature repeatedly identifies clerical and routine office tasks as highly exposed to generative AI. Draft for Commissioner Review Page 9 Austin EPC | AI Policy Recommendation | DRAFT Customer support and other structured information-handling work These jobs often involve repetitive text, triage, search, and communication tasks that current systems can assist with or partially automate. Business, finance, legal, marketing, analysis, and related knowledge work These fields often contain high-value cognitive tasks that AI can support. The main question is not only whether jobs disappear, but how roles, staffing patterns, and hiring standards change. Software, data, and technical support work Austin has a large local base here. These workers may benefit from productivity gains and demand growth, but they are also among the earliest users of AI tools. Lower-buffer households in exposed occupations Residents with less savings, less schedule flexibility, or fewer retraining options may feel labor-market change more acutely even if headline unemployment remains low. Early career workers and recent graduates If firms raise productivity expectations or slow entry-level hiring, younger workers may face weaker entry paths even when experienced workers remain employed. D. Job-Creation and Expansion Pathways Potential local job-creation or job-expansion pathways include: • AI-related technical roles. Austin’s large local base in software, data, analytics, and systems work may help the region attract and retain firms building or deploying AI products. • Implementation, cybersecurity, data governance, and training work. As organizations adopt AI, they often need staff to integrate tools, manage risk, train users, and maintain controls. • City and public-sector readiness roles. Resolution fifty-five already signals that oversight, public engagement, digital equity, and workforce considerations require staff capacity. • Construction and skilled trades tied to digital infrastructure. Data centers and related electrical or cooling infrastructure can create demand for electricians, mechanical trades, controls, network installation, and maintenance work. • Adjacent local spillovers. Large infrastructure projects and expanding technical operations can create secondary demand in local suppliers, logistics, maintenance, and business services. These pathways are real enough to track but should not be overstated. The open question is how much of that Austin residents can capture. E. Measurement Plan The initial measurement system should be simple and focused rather than comprehensive. Metric What It Tracks Suggested Source Total non-farm employment and unemployment Broad labor-market conditions BLS / TWC Frequency Monthly Draft for Commissioner Review Page 10 Occupational mix by major group Exposure-relevant job structure BLS OEWS Annual Austin EPC | AI Policy Recommendation | DRAFT Targeted occupations Wage trends in selected occupations Claims and separations by industry or occupation Hiring and postings in exposed occupations Training enrollment and completion AI-related infrastructure indicators Office support, customer support, business and financial, computer and mathematical, legal, media, and selected service roles Whether exposed work is seeing pressure or premium Early stress signal BLS OEWS / TWC Annual or quarterly where possible BLS OEWS / QCEW Annual or quarterly where possible TWC / Workforce system partners Monthly Early slowdown signal TWC and, if available, commercial postings data Monthly Local adaptation capacity Construction and skilled- trades demand ACC, Workforce Solutions Capital Area, City programs Austin Energy, permitting data, project announcements Quarterly Quarterly 5.4 Feasibility Constraints • Authority: EPC is advisory. Any programmatic response will require Council direction, staff ownership, or both. • Geography: City resident data and metro labor-market data are not identical. The brief should be explicit when it is using City of Austin figures versus Austin metro figures. • Attribution: Isolating AI from the business cycle, trade, interest rates, or firm-level restructuring is difficult. The city should monitor signals, not claim direct causation too quickly. • Data lag: Some of the best public data are annual and backward-looking. This makes a small set of timely monthly indicators important. • Capacity: Cross-department coordination will be required. This work touches Economic Development, the Innovation Office, Austin Energy, Procurement, workforce partners, and outside institutions. • Budget discipline: The first phase should rely on existing public data and existing institutional partnerships. New spending should follow only after the City has a clearer picture of the labor issue. Draft for Commissioner Review Page 11 Austin EPC | AI Policy Recommendation | DRAFT 5.5 Policy and Program Options These options are not mutually exclusive. Option A: Austin AI Labor Market Dashboard • Action: Direct relevant City staff and workforce partners to publish a small quarterly dashboard focused on AI-relevant labor indicators. • Purpose: Give Council and the public a shared evidence base. • Implementers: Economic Development Department, Innovation Office, Workforce Solutions Capital Area, Austin Energy, and research partners. • Expected effect: Better visibility into both downside risk and upside opportunity. • Resourcing: Low. Option B: Targeted Worker Transition and Skills Plan • Action: Align existing workforce and training partners around a narrow set of exposed occupations and a narrow set of growth areas. • Priority groups: Administrative support workers, customer-facing structured-information roles, early-career knowledge workers, and residents with weaker retraining access. • Growth-side alignment: Computer support, data work, cybersecurity, project implementation, and skilled trades linked to electrical and cooling infrastructure. • Implementers: Workforce Solutions Capital Area, Austin Community College, City workforce programs, public libraries, and employer partners. • Expected effect: Better resident adaptation capacity. • Resourcing: Medium. Option C: AI Infrastructure and Skilled-Trades Readiness Initiative • Action: Track AI-related infrastructure and data-center buildout in the region and connect that pipeline to local skilled-trades pathways. • Focus areas: Electricians, HVAC and cooling, controls, fiber and networking, maintenance, and related apprenticeships. • Purpose: Improve the chance that residents capture construction and maintenance upside from infrastructure growth. • Implementers: City staff, Austin Energy, ACC, contractor and apprenticeship partners, and regional workforce entities. • Expected effect: Moderate positive for local job access in selected trades. • Resourcing: Medium. Draft for Commissioner Review Page 12 Austin EPC | AI Policy Recommendation | DRAFT Option D: City AI Adoption Workforce Review • Action: For significant City AI procurements or pilots, require a short workforce review that addresses training, role redesign, accountability, public communication, and implementation support. • Purpose: Prepare the City’s own workforce and make local policy more credible. • Link to current policy: This is consistent with the City’s existing focus on oversight, workforce considerations, and public engagement. • Implementers: Innovation Office, Procurement, Human Resources, departmental owners, and Law. • Expected effect: Better internal preparedness and more disciplined deployment. • Resourcing: Low to medium. 5.6 Key Risks and Mitigations Risk Category Likelihood Mitigation Overstating AI’s near- term labor effects and designing policy around weak evidence Missing uneven effects on residents with lower financial buffers or weaker retraining access Failing to watch for entry-level hiring slowdown Focusing only on downside and missing realistic upside Scope creep across multiple AI topics # R1 R2 R3 R4 R5 R6 Analytical Medium Equity High Use a staged approach. Separate exposure, hiring trends, and actual unemployment outcomes. Pair exposure analysis with economic- security indicators such as poverty, insurance coverage, and access to training. Labor-market Medium Monitor younger-worker hiring and first-job pathways in exposed occupations. Economic development Medium Process Medium Track technical hiring, infrastructure activity, and skilled-trades demand alongside disruption indicators. Keep the brief focused on labor, economic security, and city preparedness for residents. Prefer Census, BLS, TWC, City sources, and named research institutions. Remove unsupported estimates. Relying on stale or non- authoritative claims Evidence quality High Draft for Commissioner Review Page 13 Austin EPC | AI Policy Recommendation | DRAFT 6. Stakeholder Signals and Map (External) This section captures preliminary stakeholder signals and identifies outside organizations relevant for external buy-in. The section is organized by stakeholder type rather than individual names and titles, which makes the document easier to maintain and improve over time. Organization / Stakeholder Type Workforce Solutions Capital Area Why It Matters Suggested Ask Regional workforce board with visibility into labor demand, training funds, and worker services Which occupations appear most exposed locally, and where are transition services strongest or weakest? Austin Community College Major provider of credentialed and non- credit workforce training Texas Workforce Commission State labor-market and claims data source UT Austin and other local research institutions Research supports labor economics, technology, and public policy Which short-cycle programs could support both AI-exposed workers and AI-related growth sectors? Which monthly indicators could support early warning without over-interpreting noise? Can published exposure frameworks be translated into a modest Austin-specific analysis? Austin Chamber and employer networks Visibility into local hiring, investment, and business adoption patterns Which occupations are changing first, and where are firms still hiring? Labor and worker advocacy organizations Ground-level view of work redesign, displacement concerns, and access barriers Which residents may face the most strain if AI changes work faster than training pathways can respond? Regional contractors, apprenticeship sponsors, and trades partners Practical view of infrastructure and construction demand Which skilled trades may see demand growth from digital infrastructure and utility buildout? Draft for Commissioner Review Page 14 Austin EPC | AI Policy Recommendation | DRAFT 7. Recommended Next Steps (60-Day Plan) 7.1 Days 1–15: Baseline Assembly • Confirm the working group and scope. • Build a simple baseline using City resident indicators and Austin metro labor indicators. • Select a brief list of 8–10 core metrics for recurring review. • Confirm internal departmental points of contact for labor, AI governance, procurement, and infrastructure. 7.2 Days 15–40: Impact Estimation and Stakeholder Input • Review published AI exposure and adaptation research and translated it into an Austin lens. • Conduct a small number of targeted discovery conversations with workforce, education, employer, labor, and research stakeholders. • • Identify the local occupations and resident groups that appear most important to watch. Identify realistic job-creation pathways that the city should monitor alongside downside risk. 7.3 Days 40–60: Policy Design and Synthesis • Produce a short synthesis memo with three parts: baseline, impact estimation, and policy options. • Distinguish clearly between verified facts, preliminary judgments, and open questions. • Identify which options can be launched with existing resources and which would need Council action or future funding. • Present the revised concept brief to the full Commission for direction on the next phase. • If the Commission agrees with the concept that warrants further work, advance the brief to the Exploratory phase under the Commission’s formal workflow. Draft for Commissioner Review Page 15 Austin EPC | AI Policy Recommendation | DRAFT 8. References 1. U.S. Census Bureau, QuickFacts: Austin, Texas. 2. U.S. Bureau of Labor Statistics, Austin-Round Rock-San Marcos, TX Economy at a Glance. 3. U.S. Bureau of Labor Statistics, Occupational Employment and Wages in Austin-Round Rock-San Marcos, May 2024. 4. City of Austin, April 24, 2025, Council Meeting, Item 55. 5. Austin Energy, Resource, Generation and Climate Protection Plan to 2035 and related planning materials. 6. OpenAI, GPTs are GPTs: An early look at the labor market impact potential of large language models. 7. Anthropic Labor market impacts of AI: A new measure and early evidence. 8. Brookings, Measuring U.S. workers’ capacity to adapt to AI-driven job displacement. 9. Brookings, the geography of generative AI’s workforce impacts will differ from those of previous technologies. 10. Felten, Raj, and Seamans, Occupational, industry, and geographic exposure to artificial intelligence: A novel dataset and its potential uses. Draft for Commissioner Review Page 16