2nd/10_Wiki/Topic_Programming/AI_and_ML/Activism.md

id, title, category, status, canonical_id, aliases, duplicate_of, source_trust_level, confidence_score, verification_status, tags, raw_sources, last_reinforced, github_commit, tech_stack

id	title	category	status	canonical_id	aliases	duplicate_of	source_trust_level	confidence_score	verification_status	tags	raw_sources	last_reinforced	github_commit	tech_stack
wiki-2026-0508-activism	Activism	10_Wiki/Topics	verified	self	Advocacy Social Movement Civic Engagement Tech Activism	none	B	0.85	applied	activism ethics social-movement civic-tech ai-ethics		2026-05-10	pending	language framework N/A civic-tech tooling

Activism

매 한 줄

"매 organized action toward social/political change". Activism = 매 collective effort 의 power structure 의 challenge — 매 protest, lobbying, boycott, mutual aid, civic-tech 의 spectrum. 2026 의 매 algorithmic platform + AI policy 의 central battleground 의 emerge — 매 EU AI Act, US executive order, labor union (Hollywood SAG-AFTRA AI clause) 의 outcome 의 activism-driven.

매 핵심

매 tactic spectrum

• Institutional: lobbying, policy comment, litigation (EFF, ACLU model).
• Direct action: protest, strike, blockade, boycott.
• Mutual aid: 매 community 의 direct support — 매 state 의 bypass.
• Cultural: art, media, narrative shift.
• Digital / civic-tech: open data, FOIA tooling, mapping, OSINT.

매 movement lifecycle (Stages of a Social Movement, Blumer 1969)

1. Emergence — 매 unrest, scattered grievance.
2. Coalescence — 매 leadership + identity.
3. Bureaucratization — 매 formal org.
4. Decline — 매 success / repression / co-optation / mainstream.

매 tech-activism domain (2026 hot)

• AI ethics: bias audit, dataset transparency, algorithmic accountability.
• Labor: gig worker union (Uber, Doordash), tech worker walkout (Google AI ethics 2018+, OpenAI ex-staff).
• Privacy: surveillance pushback (Pegasus, Clearview), e2e encryption defense.
• Climate tech: Tech Won't Build It, Stop Cop City.
• Open source: copyleft, fair-source debate, AI-output licensing.
• Disinformation: fact-check infra, election integrity.

매 응용

1. Coalition building (cross-org alliance).
2. Narrative campaign (framing, messaging).
3. Policy intervention (model regulation comment, e.g. NIST AI RMF).
4. Whistleblowing infra (SecureDrop, Signal).
5. Crisis mapping (Ushahidi, Bellingcat OSINT).

💻 패턴

Secure communication (org-internal)

# 매 Signal — 매 default for activist coordination
# 매 disappearing message + safety number verify + screen-lock
# 매 Wire / Element (Matrix) 의 self-host alternative

FOIA / public records request (US)

# muckrock-style template
import requests
req = {
    "agency_id": 123,
    "title": "Records re: facial recognition vendor contracts 2024-2026",
    "documents_requested": "All contracts, MOUs, RFPs related to FRT...",
    "fee_waiver_request": True,  # public interest
}
# muckrock.com API 의 submit 또는 direct email per agency

OSINT verification (geolocation)

# 매 image metadata + reverse geocode + sun-shadow angle
from PIL import Image
from PIL.ExifTags import GPSTAGS, TAGS
img = Image.open("photo.jpg")
exif = img._getexif() or {}
gps = next((v for t, v in exif.items() if TAGS.get(t) == "GPSInfo"), {})
# 매 cross-ref Sentinel-2 satellite imagery + OpenStreetMap

Mutual aid request board (simple)

# Flask + sqlite — 매 community fridge / ride share / bail fund
@app.post("/request")
def post_request():
    db.execute("INSERT INTO need (kind, area, contact, ts) VALUES (?, ?, ?, ?)",
               (request.form["kind"], request.form["area"], hashed(request.form["contact"]), now()))

Algorithmic audit (bias)

from sklearn.metrics import confusion_matrix
def disparate_impact(y_pred, y_true, group):
    rates = {}
    for g in set(group):
        mask = group == g
        rates[g] = (y_pred[mask] == 1).mean()
    return min(rates.values()) / max(rates.values())  # 매 <0.8 의 80% rule violation

Petition / call-tool (5Calls-style)

// 매 lookup rep by zip → display script → log call
const reps = await fetch(`/api/reps?zip=${zip}`).then(r => r.json());
// reps: [{ name, phone, party, district }]

매 결정 기준

상황	Tactic
Policy window open (bill in committee)	Lobbying + public comment + coalition letter
Corporate harm, public-facing	Boycott + media campaign + shareholder resolution
Acute injustice (police violence)	Protest + legal observers + bail fund
Long-term cultural shift	Narrative / art / education
Tech worker concern	Internal organizing + ethical board + walkout (last resort)
State surveillance	Encryption advocacy + legal (EFF model)

기본값: 매 multi-tactic + 매 sustained (movement >> moment) + 매 affected community 의 leadership.

🔗 Graph

• 부모: Civic Engagement
• 응용: Algorithmic Accountability
• Adjacent: AI Ethics

🤖 LLM 활용

언제: 매 policy doc summarization, 매 outreach draft, 매 large-corpus FOIA review. 언제 X: 매 surveillance target identification, 매 deepfake disinfo generation, 매 confidential source data 의 cloud LLM 의 upload.

❌ 안티패턴

• Slacktivism: 매 hashtag 의 only — 매 sustained organizing 의 X.
• Voluntourism: 매 affected community 의 lead-not 의 outsider 의 dominate.
• Burnout culture: 매 unsustainable pace — 매 movement 의 collapse.
• OPSEC fail: 매 plain-text channel, 매 metadata leak — 매 source 의 endanger.
• Performative ally: 매 statement 의 only without resource shift.
• Astroturfing: 매 fake grassroots — 매 trust 의 destroy 의 long-term.
• Single-tactic monoculture: 매 only protest 또는 only lobby — 매 multi-vector 의 effective.

🧪 검증 / 중복

• Verified (Blumer 1969 Collective Behavior; Tarrow Power in Movement 3rd ed; EFF / Access Now annual reports; AI Now Institute reports 2018-2025).
• 신뢰도 B (매 normative + tactical mix — context-dependent).

🕓 Changelog

날짜	변경
2026-05-08	Phase 1
2026-05-10	Manual cleanup — 2026 tech-activism focus + civic-tech patterns