2nd/10_Wiki/Topics/AI_and_ML/Universal Basic Income (UBI).md

---
id: wiki-2026-0508-universal-basic-income-ubi
title: Universal Basic Income (UBI)
category: 10_Wiki/Topics
status: verified
canonical_id: self
aliases: [UBI, Basic Income, Citizen's Income]
duplicate_of: none
source_trust_level: B
confidence_score: 0.8
verification_status: applied
tags: [policy, ai-economy, ubi, automation, future-of-work]
raw_sources: []
last_reinforced: 2026-05-10
github_commit: pending
tech_stack:
  language: na
  framework: policy-economics
---

# Universal Basic Income (UBI)

## 매 한 줄
> **"매 unconditional cash transfer to every citizen, regardless of work or means"**. 매 18세기 Thomas Paine 부터 Friedman의 negative income tax 까지 historical roots. 매 2020s LLM-driven labor displacement 이 debate를 mainstream으로 — Sam Altman (OpenAI), Andrew Yang, Yuval Harari 의 advocacy. 매 2026 현재 GiveDirectly Kenya, Stockton SEED, Sam Altman의 Worldcoin / OpenResearch 등 실증 pilot data 축적 중.

## 매 핵심

### 매 4 defining features
1. **Universal**: 매 citizen 모두 (means-test 없음).
2. **Unconditional**: work requirement 없음.
3. **Cash**: in-kind voucher 가 아닌 현금.
4. **Regular**: monthly / quarterly recurring (one-shot 이 아님).

### 매 motivation 의 spectrum
- **Left**: poverty elimination, labor decommodification.
- **Libertarian** (Friedman, Hayek): replace welfare bureaucracy with simple transfer.
- **Tech / AI**: automation displacement → redistribute productivity gains.
- **Feminist**: unpaid care work 의 recognition.

### 매 funding mechanisms
1. **Carbon tax dividend** (Alaska Permanent Fund model).
2. **VAT** (Andrew Yang's "Freedom Dividend" proposal).
3. **AI / data dividend** (compute tax, data tax).
4. **Land value tax** (Georgist).
5. **Sovereign wealth fund returns**.

### 매 evidence (modern pilots)
- **Finland 2017-18**: 2000 unemployed, €560/mo. wellbeing↑, employment ~neutral.
- **GiveDirectly Kenya 2017+**: 12-year RCT, 20K+ recipients. asset/business creation↑.
- **Stockton SEED 2019-21**: $500/mo. full-time employment↑ (28% → 40%).
- **Sam Altman OpenResearch 2020-23**: $1000/mo, 3 years, 3000 people in TX/IL.
- **Wales basic income for care leavers 2022+**: £1600/mo for 18-yr-olds out of foster care.

### 매 AI-induced unemployment debate
- **Optimist (Karl Benz argument)**: new jobs emerge (jevon's paradox 의 extension).
- **Pessimist (Acemoglu, Hinton)**: cognitive automation 은 historical pattern 다름.
- **Empirical 2024-26**: software engineering, customer service, translation 의 measurable wage compression. 그러나 net unemployment 는 not yet structural.

### 매 응용
1. Poverty floor.
2. Bargaining power for low-wage workers (exit option).
3. Entrepreneurship enablement.
4. Caregiver / artist subsidy.
5. AI displacement insurance.

## 💻 패턴 (policy design)

### Pattern 1: Negative Income Tax (Friedman)
```text
benefit = max(0, (threshold - income) × rate)
# threshold = $30K, rate = 50%
# income $0   → benefit $15K
# income $20K → benefit $5K
# income $30K → benefit $0
# Phase-out automatic, but not "universal".
```

### Pattern 2: Pure UBI flat
```text
benefit = $1000/month for every adult citizen
funding = 10% VAT + carbon dividend
# Universal, but expensive. ~$3T/yr in US.
```

### Pattern 3: Alaska Permanent Fund (real-world precedent)
```text
oil_revenue → sovereign_fund (1976-)
annual_dividend = fund_5yr_avg_return × payout_ratio
# 1982-2024: $1000-$3200 per resident annually.
# 매 only true unconditional cash transfer at scale 의 sustained example.
```

### Pattern 4: AI / compute dividend (proposed, 2020s)
```text
tax = compute_used × rate  // foundation model training
fund = sovereign_AI_fund
dividend = fund_return / population
# Not yet implemented. Sam Altman의 OpenAI public benefit proposal.
```

### Pattern 5: Pilot RCT design
```text
1. Random assignment: treatment ($500-$1000/mo) vs control.
2. Duration: ≥3 years (recover Hawthorne, capture habit formation).
3. Outcomes: employment, wellbeing, health, education, household formation.
4. Pre-registration: avoid p-hacking.
5. Heterogeneity analysis: by income, age, family structure.
```

## 매 결정 기준
| 정책 목표 | Approach |
|---|---|
| Poverty elimination only | Targeted transfer / NIT |
| Reduce welfare bureaucracy | NIT or UBI replacing means-tested |
| AI displacement hedge | UBI + retraining stipend |
| Resource curse (oil, gas) | Sovereign fund dividend (Alaska) |
| Care work recognition | Targeted caregiver UBI |

**기본값**: NIT for fiscal pragmatism, full UBI for radical simplicity (high cost).

## 🔗 Graph
- 부모: [[Welfare Policy]] · [[AI and Economics]]
- 변형: [[Negative Income Tax]] · [[Job Guarantee]] · [[Sovereign Wealth Dividend]]
- 응용: [[Poverty Alleviation]] · [[Future of Work]] · [[AI Safety Policy]]
- Adjacent: [[Automation]] · [[Technological Unemployment]] · [[Yang Freedom Dividend]]

## 🤖 LLM 활용
**언제**: AI economic impact 분석, automation policy 토론, welfare reform 비교.
**언제 X**: pure macro forecasting (UBI 의 macro effect 의 unsettled).

## ❌ 안티패턴
- **"UBI fixes everything"**: housing / healthcare 의 supply constraint 는 cash transfer 로 해결 X.
- **Confounding NIT with UBI**: means-tested 와 universal 의 혼동.
- **Ignoring inflation**: $1000/mo 이 housing-supply-constrained 시장에선 rent 로 흡수.
- **Cherry-picking pilot results**: short pilot 은 long-term 행동 변화 capture 못 함.
- **Replacing all welfare**: disability, healthcare 의 specialized support 까지 cash 로 대체 시 vulnerable 그룹 손해.

## 🧪 검증 / 중복
- Verified (OpenResearch 2024 final report; GiveDirectly RCT 2024; Stockton SEED 2-yr report 2021; Hoynes & Rothstein review NBER 2019).
- 신뢰도 B (정책 영역, contested empirical evidence).

## 🕓 Changelog
| 날짜 | 변경 |
|---|---|
| 2026-05-08 | Phase 1 |
| 2026-05-10 | Manual cleanup — UBI definition, pilot data, AI displacement context |