AI summary by TwinMind
5 Ethical Principles for Digitizing Humanitarian Aid | Aarathi Krishnan | TED
TED
11m 46s
Opening Context and Rohingya Case Study
- Zeynep Tufekci quote: history full of harm from powerful people with good intentions
- 2017: Rohingya refugees fled Myanmar military crackdown (UN called genocidal intent)
- Refugees forced to register for government-backed digital biometric ID cards in Bangladesh camps
- No opt-out option provided
- 2021: Human Rights Watch accused humanitarian agencies of improperly sharing data with Myanmar government
- Data included biometrics, family makeup, overseas relatives, origin locations
- Sparked retaliation fears, some refugees went into hiding
- Digital data now faster to access, quicker to scale, more readily accessible than historical targeting methods
- Failure on institutional, governance, and moral fronts
Speaker's Background and Humanitarian Tech Acceleration
- 15 years humanitarian aid career (Rwanda to Afghanistan)
- Humanitarian aid definition: emergency care provision (food, water, shelter) post-disaster/crisis
- Experience: led multicountry global programs, designed drone innovations for disaster management across small island states
- Worked with communities in fragile contexts on future planning
- Designed global strategies for humanitarian organizations
- Observation: humanitarians embraced digitalization at incredible speed over last decade
- From tents/water cans to AI, big data, drones, biometrics
- Deployment of untested technologies on vulnerable populations without appropriate consent
Colonial Nature of Humanitarian Technology
- Current global agonies result from shared colonial history
- Humanitarian tech innovations inherently colonial:
- Designed for groups seen as outside technology
- Recipients not recognized as capable of providing own solutions
- Core question: How to avoid locking people into future harm, indebtedness, inequity?
- Personal motivation: people from speaker's historically excluded/marginalized communities often denied voice in future choices
21st Century Challenges and Digital Futures as "Balm"
- Future characterized by: deep uncertainty, dying planet, distrust, pain
- Digital futures viewed as soothing balm, logical inevitability
- Recent reports flag new technology risks:
- Data collected on vulnerable individuals used for retaliation
- Risk extends to families and communities
Afghanistan Case Study (August 2021)
- Taliban takeover resulted in biometric data collected by US military/Afghan government falling into Taliban hands
- Journalists' houses searched
- Afghans raced to erase digital history online
- Technologies of empowerment became technologies of disempowerment
Technology Design Problems
- Technologies designed on specific societal assumptions
- Embedded in market, filtered through capitalist considerations
- Context-switching failure: technologies created in one context fail when parachuted into another
- Example: fingertip scan comfort for movies vs. safety concerns for food rations access
- Humanitarians assume technology liberates without considering power, exploitation, harm
- "Solutionizing" = magical thinking that shiny solutions solve problems without analyzing underlying realities
- Tools analogy: chef's knife creates beautiful meal or devastation depending on user
Positive Examples and Way Forward
- Indigenous AI: draws from Indigenous protocols/values instead of Western philosophies for AI code
- Nia Tero: Indigenous co-led organization enabling communities to map own well-being/territories
- Satellite Sentinel Project (2010): mapped atrocities through remote sensing to predict/prevent
- Wound down after few years due to inability to generate action
- More importantly: team realized operating without ethical net
- Decided to wind down before creating harm
Ethical Principles Framework
- Ask: Which groups will be harmed and when?
- Assess: Who actually benefits from this solution?
- Interrogate: Was appropriate consent obtained from end users?
- Consider: What must we gracefully exit to be fit for futures?
- Imagine: What future good might we foreclose if implemented today?
Conclusion
- Accountability for futures we create
- Cannot absolve responsibilities if actions cause harm to those we serve
- "Another world is absolutely, radically possible"
- Position: "I'm not anti-tech. I am anti-dumb tech"
