Strengthen America A 21st-Century Compact

§ Legislative Act

Workforce Automation Transition and Algorithmic Accountability Act (WATAA)

Current Status

Existing Law: Worker Adjustment and Retraining Notification (WARN) Act (29 U.S.C. § 2101-2109) requires 60-day notice for mass layoffs of 100+ workers. Trade Adjustment Assistance (TAA) (19 U.S.C. § 2271) covers trade-displaced workers only. No federal statute addresses automation-specific displacement or algorithmic employment decisions.

Current Authority: Department of Labor administers unemployment insurance (UI) and WARN enforcement. EEOC handles employment discrimination but lacks explicit algorithmic audit authority. States administer UI with federal oversight.

Existing Limitations: WARN applies only to "plant closings" and "mass layoffs"—gradual automation attrition (e.g., 5% annually) escapes coverage. TAA excludes technology-displaced workers entirely. No federal standard governs algorithmic termination, performance scoring, or worker data rights. UI benefit duration (26 weeks average) insufficient for mid-career retraining requiring 12-18 months.

Problem

Specific Harm: McKinsey Global Institute projects 12 million occupational transitions required by 2030 in the United States due to automation and AI adoption.¹ Bureau of Labor Statistics data shows displaced workers aged 45+ experience 23% permanent wage reduction upon reemployment.² Current retraining programs reach <5% of displaced workers. Algorithmic management systems now cover 60%+ of hourly workers, with no transparency requirements or appeal rights.³

Who is Affected: Manufacturing, retail, transportation, and clerical workers—disproportionately non-college-educated workers aged 40-60 in regions with concentrated industry exposure (Rust Belt, rural communities). Secondary effects on municipal tax bases and healthcare systems in high-displacement areas.

Gaps in Current Law: (1) No trigger mechanism for gradual automation displacement. (2) No portable, pre-funded retraining infrastructure. (3) No algorithmic transparency or due process standards. (4) No revenue mechanism linking automation productivity gains to worker transition costs. (5) WARN enforcement relies on private litigation with no proactive agency monitoring.

Accountability Failures: Workers terminated by algorithmic systems currently appeal to the same employer's HR department that deployed the algorithm. No independent review exists. EEOC lacks technical capacity to audit algorithmic discrimination. Department of Labor lacks authority to require automation impact disclosures. GAO has documented that existing workforce programs operate in silos with no outcome tracking across agencies.4

Proposed Reform

Primary Policy Change: Create a unified federal Automation Displacement Response System combining (1) extended income support tied to mandatory retraining participation, (2) employer-funded portable retraining accounts with federal matching, and (3) binding algorithmic transparency and independent appeal rights. Funded from general revenue as part of comprehensive federal reform.

New Requirements: Employers with 100+ workers must provide 90-day Automation Impact Notices when technology adoption reduces headcount by 10%+. Firms using algorithmic management for employment decisions must submit to annual third-party audits. All algorithmic termination, demotion, or discipline decisions require human review and independent appeal pathway via GAO Labor Docket with FTC algorithmic oversight. Employers contribute $500/year per worker to portable Universal Retraining Accounts.

New Prohibitions: Termination, demotion, or material discipline based solely on algorithmic output without human review. Retaliation against workers exercising algorithmic appeal rights. Use of algorithmic management systems that fail third-party disparate impact audits without remediation within 180 days.

Enforcement: Department of Labor gains proactive WARN+ monitoring authority with subpoena power. GAO Labor Docket with FTC algorithmic oversight provides binding arbitration authority over algorithmic employment disputes. Civil penalties up to $50,000 per violation for notice failures, $100,000 per violation for algorithmic accountability breaches.

Definitions

"Algorithmic management system": Any computational system, including artificial intelligence, machine learning, or automated decision-making tools, that collects worker data, evaluates worker performance, makes recommendations regarding employment decisions, or makes or substantially influences decisions regarding scheduling, compensation, discipline, demotion, promotion, or termination.

"Automation": The use of technology, machinery, software, robotics, artificial intelligence, or other non-human systems to perform tasks previously performed by human workers, including but not limited to physical automation of manufacturing and logistics, software automation of clerical and administrative functions, and AI automation of analytical, creative, or decision-making functions.

"Automation-displaced worker": A worker certified as having been separated from employment due to automation, entitled to extended unemployment benefits and enhanced retraining access under this Act.

"Certified training provider": An educational institution, apprenticeship program, industry training provider, or online learning platform that meets quality standards and is approved by DOL to accept Universal Retraining Account payments.

"Covered worker": For purposes of Universal Retraining Accounts, any employee who has completed 90 days of employment with an employer subject to contribution requirements.

"Disparate impact": For purposes of algorithmic audit requirements, a statistically significant difference in outcomes by protected class consistent with standards established in Griggs v. Duke Power Co., 401 U.S. 424 (1971), and EEOC Uniform Guidelines on Employee Selection Procedures.5

"Full-time equivalent employee": For purposes of employer thresholds, calculated by dividing total hours worked by all employees during the preceding 12 months by 2,080.

"GAO Labor Docket": The specialized docket within the GAO providing independent worker appeals, dispute resolution, and binding arbitration for algorithmic employment decisions, with FTC providing technical algorithmic oversight and audit certification.

"Human review": For purposes of algorithmic accountability requirements, review and decision-making by a natural person with: (i) authority to override algorithmic recommendations, (ii) access to underlying data and algorithmic assessment, (iii) accountability for the employment decision, and (iv) documentation of independent judgment exercise.

"Universal Retraining Account (URA)": A portable, individually-owned account for the purpose of funding worker education, retraining, and skill development.

What Changes

Before: WARN Act covers only mass layoffs of 100+ workers with 60-day notice. Does not address gradual automation attrition or technology-specific displacement. No algorithmic employment accountability. Workers terminated by algorithm appeal only to employer. No portable retraining funding. No revenue mechanism connecting automation gains to transition costs. Displaced workers receive 26 weeks UI only. No federal training quality standards with enforcement. Community impact unaddressed.

After: WARN+ expands coverage to 10%+ workforce reductions from automation with 90-day notice. Universal Retraining Accounts provide $5,000 portable funding per worker refreshed every 5 years. Extended UI of 78 weeks for automation-displaced workers in retraining. GAO Labor Docket with FTC algorithmic oversight adjudicates algorithmic employment disputes with binding authority. Mandatory human review for algorithmic termination/discipline decisions. Annual algorithmic audits for large employers with FTC-certified auditors. $10B annual Community Transition Grants. Worker data rights with portability. Pilot programs for UBI, job guarantee, and 4-day week with independent GAO evaluation. GAO biennial oversight of all program components. Funded from general federal revenue.

ROI

Costs:

Item 10-Year
Displacement insurance (wage insurance + extended UI) $230,000,000,000-$320,000,000,000
Universal Retraining Accounts (federal match) $250,000,000,000
Training infrastructure investment $250,000,000,000
Community Transition Grants $100,000,000,000
GAO Labor Docket and program administration $50,000,000,000-$70,000,000,000
Total $880,000,000,000-$990,000,000,000

Savings:

Item Gross Capture Net
Employer contributions to URAs $750,000,000,000 100% $750,000,000,000
Reduced standard UI expenditure $50,000,000,000-$100,000,000,000 100% $50,000,000,000-$100,000,000,000
Reduced SNAP, Medicaid, other assistance $100,000,000,000-$150,000,000,000 80% $80,000,000,000-$120,000,000,000
Total $880,000,000,000-$970,000,000,000

Societal Benefits:

Benefit Annual NPV (3%) NPV (7%)
Avoided permanent wage scarring $15,000,000,000-$25,000,000,000 $120,000,000,000-$200,000,000,000 $90,000,000,000-$150,000,000,000
Preserved municipal tax bases Not quantified Not quantified Not quantified
Reduced litigation and EEOC backlog $500,000,000 $4,000,000,000 $3,000,000,000
Long-term productivity gains $50,000,000,000 $400,000,000,000 $300,000,000,000

Summary:

Category 10-Year Notes
Net federal cost $630,000,000,000-$840,000,000,000 Funded from general revenue
Worker reemployment improvement 33% reduction in displacement time From 27 to 18 weeks average
Post-displacement wage preservation From 77% to 90% Target wage replacement rate
Training completion rate 60%+ Versus current ~40%
Automation notice compliance 95%+ Versus current WARN ~80%

Funding Note: Program costs funded from general federal revenue. Comprehensive revenue reform generates $852B annual surplus, of which workforce transition represents approximately 8-10%. See Revenue Model Overview (0_Revenue_Model_Overview.md) for federal fiscal framework.

References

  1. McKinsey Global Institute, "Jobs Lost, Jobs Gained: Workforce Transitions in a Time of Automation" (2017, updated 2021)
  2. Bureau of Labor Statistics Displaced Worker Survey (2022)
  3. UC Berkeley Labor Center, "Data and Algorithms at Work: The Case for Worker Technology Rights" (2023)
  4. GAO-22-104581, "Workforce Programs: DOL Could Do More to Ensure States Collect and Report Complete Information" (2022)
  5. Griggs v. Duke Power Co., 401 U.S. 424 (1971); EEOC Uniform Guidelines on Employee Selection Procedures
  6. Brookings Institution, "Automation and Artificial Intelligence: How Machines Are Affecting People and Places" (2019)
  7. Worker Adjustment and Retraining Notification Act, 29 U.S.C. § 2101-2109
  8. Trade Adjustment Assistance, 19 U.S.C. § 2271 et seq.
  9. Workforce Innovation and Opportunity Act, 29 U.S.C. § 3101 et seq.
  10. Title VII of Civil Rights Act, 42 U.S.C. § 2000e
  11. Age Discrimination in Employment Act, 29 U.S.C. § 621 et seq.
  12. GAO-23-105597, "Artificial Intelligence: An Accountability Framework for Federal Agencies and Other Entities" (2021)
  13. Congressional Research Service R46795, "Worker Displacement and Adjustment Assistance" (2021)
  14. German Kurzarbeit (short-time work) program structure
  15. EU AI Act algorithmic accountability provisions (Regulation 2024/1689)
  16. Singapore SkillsFuture portable training account model
  17. Wards Cove Packing Co. v. Atonio, 490 U.S. 642 (1989)
  18. International Union, UAW v. Johnson Controls, 499 U.S. 187 (1991)