Skip to main contentSkip to navigationSkip to search

Get The Peoples' Hand on the AI Wheel!

The Principle

AI shapes everything - jobs, truth, power, and the environment. Decisions about AI are political decisions, not just technical ones. If only experts and executives control AI, we get techno-feudalism. Democratic control means:

  1. Workers Have the Power where AI affects their jobs
  2. Communities Have the Power where AI affects their neighborhoods
  3. The Public has the Power over AI's role in society
  4. Affected People Decide - not just those who profit

1. FAIA Democratic Governance

A. FAIA Board

Composition (15 Members Total):

  • 5 Worker Representatives - Elected by labor unions from AI-impacted industries (tech workers, displaced workers, service workers, manufacturing, and the public sector)
  • 5 Community Representatives - Selected by sortition (random selection) from communities near data centers + communities harmed by AI (deepfake victims, surveillance-impacted communities, etc.)
  • 5 Technical Experts - Computer scientists, AI safety researchers, environmental scientists, civil rights technologists, ethicists
    • NONE can have worked for Big Tech in last 10 years
    • NONE can own >$5k in tech stocks
    • NONE can receive corporate funding

Why This Composition:

  • Workers know where AI hurts workers
  • Communities know where AI hurts communities
  • Experts provide technical analysis in service of workers and communities
  • NO corporate executives - they have enough power everywhere else

Board Powers:

  • Approve/Deny All Major AI Deployments (final authority)
  • Set Enforcement Priorities (where FAIA focuses resources)
  • Override Director Decisions (2/3 vote required)
  • Approve FAIA Budget Allocation
  • Hire/fire FAIA Director (cause required, 2/3 vote)
  • Subpoena Power (investigate companies, government agencies)
  • Emergency Shutdown Authority (any 3 members can trigger immediate AI system shutdown pending Board review)

Term Limits: 4 years, one renewal maximum (prevents capture)

Compensation: $200k/year (same as senior government officials - enough to leave private sector job, not enough to get rich)

Meeting Requirements:

  • Monthly public meetings (live-streamed and archived)
  • Community hearings quarterly (rotate locations around country)
  • Emergency meetings within 24 hours of any major AI incident
B. Regional AI Councils

Structure:

  • 50 Regional Councils (one per state)
  • 15 Members Each (5 workers, 5 community, and 5 experts - same formula as national Board)

Powers:

  • Approve/Deny Data Centers in Their Region
  • Monitor Local AI Deployment (job displacement and environmental impacts)
  • Hold Public Hearings on AI impacts
  • File Complaints to the National FAIA Board
  • Community Veto - Can block any AI deployment in their region with 2/3 vote

Why Regional:

  • Data center in Arizona affects Arizona communities.
  • New York workers are displaced by AI should have say over that deployment.
  • Democracy means local control where impacts are local.
C. Sectoral Worker Councils

10 Sectoral Councils Covering:

  1. Healthcare workers
  2. Education workers
  3. Customer service/call center workers
  4. Transportation/logistics workers
  5. Manufacturing workers
  6. Creative workers (writers, artists, musicians)
  7. Legal/professional services workers
  8. Retail/hospitality workers
  9. Tech industry workers (yes, even Big Tech workers get representation)
  10. Public sector workers

Composition: Elected by workers in that sector (union and non-union)

Powers:

  • Pre-Approval Consultation - Companies MUST consult with relevant Sectoral Council before deploying AI that affects that industry
  • Job Displacement Review - Council reviews and approves/denies all retraining plans
  • Collective Bargaining over AI deployment (works with unions)
  • Recommend FAIA Enforcement Actions
  • Direct Complaints to the FAIA Board

Example: Amazon wants to deploy AI that replaces 10,000 warehouse workers → Transportation/Logistics Worker Council must approve retraining plan, compensation, and transition timeline BEFORE deployment.

2. Community Participation Mechanisms

A. Mandatory Public Hearings

Triggered by:

  • Any data center >40 megawatts
  • Any AI deployment replacing >100 workers in a community
  • Any AI system used by local government (policing, benefits, etc.)
  • Any AI with documented bias affecting protected groups
  • Community petition with 5,000 signatures

Requirements:

  • 90-Day Minimum public comment period
  • 3 Public Hearings Minimum (evening hours, childcare provided, translation services, and accessible venues)
  • Community Impact Report published 30 days before hearings (plain language, multiple languages)
  • Company Representatives Must Attend and answer questions (CEOs are required for major deployments)
  • Live Testimony Recorded and made public
  • FAIA Must Respond to every substantive comment in writing

Community veto: If 60%+ of residents in affected area oppose (measured by referendum), Regional Council can block deployment

B. Citizen AI Oversight Committees

Structure:

  • Community Members Are Selected by Sortition (like jury duty, but for AI oversight)
  • 12 Members per Committee (demographically representative of community)
  • 6-month Terms, Rotating (prevents capture, builds civic capacity)
  • Paid Stipend ($5,000 for 6 months - accessible to working people)

Responsibilities:

  • Monitor Local AI Systems (data centers, surveillance, and automated decision-making)
  • Receive Community Complaints
  • Tour Facilities Quarterly (unannounced inspections are allowed)
  • Report to Regional Council
  • Recommend Enforcement Actions

Powers:

  • Subpoena Local Records (within their jurisdiction)
  • Hold Public Hearings
  • Issue Public Reports
  • Trigger FAIA Investigations

Why Sortition: Elections can be captured by money. Random selection gives ordinary people real power. Ancient Athens used it; we can too.

C. Participatory Technology Assessment

Process:

  1. FAIA Announces a New AI Technology requiring assessment (e.g., "AI that writes code", "AI that drives trucks")
  2. Public Nominations - Communities nominate questions, concerns, priorities
  3. Citizen Panels Is Formed - 100 randomly selected citizens (demographically representative) + 50 affected workers + 20 technical experts
  4. Deliberative Process - 6 weekends of education, discussion, and debate (paid participation)
  5. Public Recommendations - Panel issues report on whether/how technology should be deployed
  6. FAIA Board Must Respond - Accept, modify, or reject with detailed explanation

Why This Matters: Instead of tech companies deploying first and apologizing later, Society Decides Collectively whether we even want this technology.

3. Worker Power in AI Development

A. Workplace AI Councils

Required at every workplace with:

  • 50+ employees AND
  • AI systems that affect work processes, scheduling, monitoring, or hiring

Composition:

  • Elected by Workers (2-year terms)
  • 7-15 Members depending on workforce size
  • Protected from Retaliation (firing a council member = $1M fine + criminal charges)

Powers:

  • Approve/deny AI Deployment in their workplace
  • Negotiate AI Terms (monitoring limits, productivity quotas, and break times)
  • Access All AI Data about workers (what's collected, how it's used)
  • Shut Down Abusive AI (majority vote triggers immediate shutdown pending FAIA review)
  • File FAIA Complaints directly

Example: Amazon warehouse wants facial recognition + productivity tracking AI → Workplace AI Council must approve, can require limits (e.g., "no bathroom break monitoring"), can veto entirely.

B. Collective Bargaining over AI

New NLRA Provisions:

  • AI Is a Mandatory Subject of Bargaining - Employers MUST negotiate over AI deployment, cannot implement unilaterally
  • Information Rights - Unions entitled to full information about AI systems, algorithms, and training data, performance metrics
  • Veto Power in Contracts - Unions can negotiate contract clauses requiring union approval for AI changes
  • Strike Protection - Workers can strike over AI deployment without losing protections

Why This Works: Unions already understand workplace power dynamics. AI bargaining becomes another tool for worker control.

4. Prevent Regulatory Capture

A. Eliminate the Revolving Door

Lifetime Bans:

  • FAIA Board Members → Can NEVER work for Big Tech
  • FAIA Director → 20-year ban on tech industry employment
  • FAIA Senior Staff → 10-year ban on tech industry employment

Why Lifetime:

  • If you can cash out later, you're incentivized to be "reasonable" now.
  • Public service = permanent commitment.

Violation Penalties:

  • Individual: Return all FAIA salary + $5M fine + criminal prosecution
  • Hiring company: $100M fine + FAIA can revoke all AI approvals
B. Financial Transparency & Conflicts

Required Disclosures (public and quarterly):

  • All assets, stock holdings, and investments
  • Family member employment in tech
  • Speaking fees, consulting income
  • Travel, gifts, and meals over $50

Automatic Recusal:

  • Any financial interest in any company being regulated
  • Family employed by the company
  • Previous employment/consulting within 10 years

Independent Auditing:

  • Senate-appointed Inspector General audits the FAIA quarterly
  • Random financial audits of Board members
  • Whistleblower protection + $5M rewards
C. Democratic Accountability Mechanisms

Annual Public Performance Review:

  • FAIA publishes report on enforcement actions, outcomes, and budget
  • Public hearings in 10 cities (community feedback)
  • Worker councils and community groups submit grades
  • Congress holds oversight hearing
  • Report cards published online

Removal Mechanisms:

  • Board Members: 2/3 Board vote OR petition of 100,000 citizens triggers recall election
  • Director: Board can remove (2/3 vote) OR Congress can impeach
  • Staff: Director can remove, but whistleblowers have appeal rights

Budget Transparency:

  • Every dollar accounted for publicly
  • Real-time spending dashboard
  • Quarterly audits (published)
  • No secret budgets, no classified AI programs are hidden from oversight

5. Balancing Expertise & Democracy

The Tension: AI regulation requires technical knowledge. But "trust the experts" = technocracy. How do we balance?

The Solution - Deliberative Democracy:

  1. Experts Advise, People Decide
    • Technical experts provide analysis, risk assessments, and options
    • Democratic bodies make final decisions based on values, priorities, and affected community needs
  2. Accessible Expertise
    • FAIA funds "people's experts" - independent scientists who explain technical issues in plain language to community groups
    • Technical training for Board members, council members (ongoing education)
    • Public libraries of accessible AI explainers
  3. Precautionary Principle
    • When experts disagree, err on the side of caution
    • Burden of proof is on companies to prove safety, not on the public to prove harm
    • "We don't fully understand this" = legitimate reason to say no
  4. Transparency Requirements
    • All technical analysis published (no hiding behind "proprietary")
    • Peer review by independent scientists
    • Community groups can hire own experts (FAIA funds this)

Example in Practice:

  • Company proposes new AI system
  • FAIA technical staff analyzes (risk assessment, bias audit, and environmental impact)
  • Report is published in technical AND plain language versions
  • Community groups/worker councils get 60 days to review and hire own experts
  • Public hearings are held
  • Democratic bodies (Board, Regional Council, and Workplace Council) make a decision based on technical analysis + community values + affected people's preferences

6. International Democratic Coordination

Problem: AI is global. US regulation matters, but can't work alone.

Solution - Democratic AI Alliance:

Structure:

  • Coalition of democracies committed to democratic AI governance
  • Each country has democratic oversight (worker councils, community boards, etc.)
  • International coordination body (delegates from each country's democratic bodies)

Goals:

  • Harmonize Standards (so companies can't jurisdiction-shop)
  • Share Enforcement (US can help EU prosecute violations, vice versa)
  • Joint Research on AI safety, detection, and impacts
  • Mutual Support for democratic movements in each country

Red Lines:

  • NO corporate representatives at the table
  • NO autocracies (China, Russia, etc. can't water down standards)
  • NO "competitiveness" arguments can be used to weaken protections

7. Enforcement of Democratic Decisions

What happens when companies ignore democratic oversight?

Escalating Penalties:

  1. Ignore Community Hearing Requirement → $10M fine + deployment blocked
  2. Deploy without Worker Council Approval → $50M fine + immediate shutdown + criminal charges for executives
  3. Circumvent Regional Council Veto → $500M fine + company banned from AI in that state for 5 years
  4. Ignore FAIA Board Decision → $5B fine + nationalization of AI infrastructure + executives are prosecuted

Criminal Penalties for Executives:

  • Undermining the Democratic Process (bribing council members, fake public comments, etc.) → 10 years in prison
  • Deploying after democratic veto → 15 years prison
  • Retaliation Against democratic Participants (firing a worker council member, threatening community organizers) → 20 years in prison

Why So Harsh: Democracy dies when powerful interests can ignore it. These penalties make clear: Democratic Decisions Are FINAL.

8. Section Summary

Multiple Layers of Democracy:

  1. National - FAIA Board sets overall policy, approves major deployments
  2. Regional - 50 Regional Councils control local AI/data centers
  3. Sectoral - 10 Worker Councils represent industries
  4. Community - Citizen oversight committees + public hearings
  5. Workplace - Worker councils control AI at job sites

No Single Point of Failure:

  • Capture the National Board? Regional Councils can still block.
  • Capture experts? Workers and communities can still veto.
  • Money floods elections? Sortition committees are immune.

Real Power:

  • Veto authority at multiple levels
  • Subpoena power, enforcement power
  • Criminal penalties for violations
  • Budget control, hiring/firing authority

Ongoing Accountability:

  • Public meetings, published reports
  • Removal mechanisms, recall elections
  • Whistleblower protections, financial transparency
  • Community grades, performance reviews