Teacher
lesson plans, instructional methods, assessments
Keith Institute Initiative
AI Is Learning From Labor. Labor Should Measure What It Contributed.
The Data Union gives labor organizations a practical method for identifying, witnessing, valuing, and bargaining over the informational assets their members create.
Labor unionOrganizes workers
Data UnionOrganizes informational assets
MethodIdentify - Witness - Value - Assert
What is a Data Union?
A Labor Institution for Workforce-Generated Information
A Data Union is not a replacement for an existing union. It is a new institutional capability that allows a labor organization to identify, document, govern, value, and assert interests relating to workforce-generated informational assets.
A labor union organizes workers.
A Data Union organizes the informational assets produced by those workers.
Nurse
clinical observations, care workflows, professional judgments
Federal employee
case decisions, administrative processes, policy interpretations
Driver
routing knowledge, delivery decisions, safety adaptations
Call-center worker
conversations, troubleshooting, escalation methods
Manufacturing worker
quality corrections, machine adjustments, maintenance knowledge
Writer/performer
creative works, voice, performance, revisions
Why now?
Artificial Intelligence Changed the Economics of Work
Historically, organizations treated worker-generated information as documentation of completed labor. AI increasingly turns that information into an input for future production.
- Worker Activity
- Informational Asset
- AI / Automation
- Productive Capability
- Economic Value
Data Union method
Identify, Witness, Value, Assert
The framework turns a broad concern about AI and work into an evidence-producing method that labor organizations can operate.
Identify
Determine what informational assets workers create.
- decisions
- observations
- communications
- methods
- creative work
- institutional knowledge
Witness
Create a verifiable record without publishing confidential underlying data.
- existence
- contributors
- chronology
- provenance
- integrity
- chain of custody
Value
Estimate contribution through multiple valuation lenses.
- cost
- market
- income
- productivity
- risk reduction
- strategic value
- Human Equivalent Work Units
Assert
Use evidence and valuation in institutional, bargaining, and legal settings.
- collective bargaining
- governance
- licensing
- regulatory engagement
- arbitration
- litigation review
- insurance review
Witnessing establishes evidence of existence, attribution, chronology, provenance, and chain of custody. Ownership and legal enforceability depend on applicable law, contracts, bargaining rights, and factual circumstances.
Human Equivalent Work Units
Measure the Work AI Can Now Perform
Keith Institute's Human Equivalent Work Unit methodology asks how much productive human-equivalent work an AI or automated system can perform because human-generated informational assets contributed to its capability.
Input valuation
What is the informational contribution worth?
HEWU
How much productive labor does the resulting system now perform?
Illustrative assessment factors
- annual task volume
- equivalent human time
- quality
- automation level
- human oversight
- workforce-generated information contribution
Data Union Action Workbook
A Workbook Designed to Produce Evidence, Not Just Discussion
The Data Union Action Workbook contains 12 modules and operational appendices with forms, playbooks, HEWU guidance, and registry specifications.
12-step process
From Governance to Annual Audit
The process is designed to move from institutional setup to asset evidence, valuation, bargaining strategy, and ongoing operation.
- 1Establish the Data Union
- 2Identify Informational Assets
- 3Map Productive Systems
- 4Witness and Register Assets
- 5Value Informational Assets
- 6Calculate Our HEWU
- 7Develop Bargaining & Legal Strategy
- 8Prepare the Demand Package
- 9Quantify Informational Labor Risk
- 10Negotiate Governance & Contract Language
- 11Operate the Data Union
- 12Conduct the Annual Audit
Examples by sector
The Informational Assets Are Already in the Work
Different occupations generate different forms of knowledge, judgment, records, methods, and creative contribution.
Government
- claims decisions
- procurement analysis
- regulatory interpretations
- case-management procedures
Education
- lesson plans
- assessment methods
- instructional strategies
- curriculum development
Healthcare
- clinical observations
- care coordination
- nursing documentation
- patient-safety interventions
Communications
- customer conversations
- troubleshooting procedures
- technical documentation
- escalation methods
Manufacturing
- quality corrections
- machine adjustments
- maintenance knowledge
- safety procedures
Transportation
- routing decisions
- logistics adjustments
- dispatch knowledge
- delivery optimization
Entertainment
- scripts
- voice
- performances
- revisions
- creative methods
Human rights + economic growth
Informational Agency Is a Human Rights Issue. Shared Productivity Is an Economic Growth Strategy.
The Data Union framework rests on two related principles: workers should have meaningful agency over how information produced through their labor is attributed, governed, reused, and commercialized; and labor should have tools to negotiate participation when AI-driven productivity uses accumulated human informational contributions.
The goal is not to stop AI adoption. The goal is to create institutions through which workers can participate in AI-driven productivity growth.
Informational Labor Risk
When Labor Can Measure the Claim, Employers Have a New Risk to Manage
Informational Labor Risk (ILR) is the possibility that organizations incur economic, contractual, regulatory, bargaining, or legal obligations arising from workforce-generated informational assets embedded in AI-enabled productive systems.
- Worker Informational Assets
- Data Union
- Evidence + Valuation
- Bargaining / Claim
- Informational Labor Risk
- AI-training compensation disputes
- data-labor claims
- unauthorized commercialization
- licensing disputes
- productivity-sharing demands
- bargaining disputes
- vendor reuse
For employers and insurers
Keith Institute develops the labor-side methodology. InclusionScore provides employer- and insurer-facing AI assurance and risk assessment for organizations managing AI-related exposure.
Witness registry
You Do Not Have to Publish the Asset to Prove It Exists
A compatible witnessing system can document identity, attribution, chronology, verification, provenance, and chain of custody while protected source material stays private.
- asset identity
- contributor class
- timestamps
- cryptographic verification
- source system
- provenance
- chain of custody
Keith Institute is exploring registry infrastructure, including platforms such as Made.CX, that can support independent witnessing and provenance while preserving confidential source material.
90-day pilot
A Union Does Not Need to Transform Overnight
A focused pilot can establish governance, identify assets, map productive systems, witness evidence, produce initial valuations, estimate HEWU, and define bargaining priorities.
Data Union Action Workbook
Build Your Data Union
The Data Union Action Workbook is a practical implementation guide for union presidents, bargaining committees, labor counsel, economists, researchers, and technology staff.