Identity for All – Refugees, Human Trafficking, Women, & Marginalized People = Tech Meets Real Life Experience & The Humans that DID + SSI Can Help Most, How & Why

From IIW

Identity For All: Refugees, Human Trafficking, Women & Marginalized Populations. Tech Meets Real Life Experience + The Humans That DID & SSI Can Help Most – How & Why.

Wednesday 7H

Convener(s): Kristina Yasuda & Jessica Hubley

Notes-taker(s): Ben Gregori & Pam Dingle

Tags for the session - technology discussed/ideas considered:

Discussion notes, key understandings, outstanding questions, observations, and, if appropriate to this discussion: action items, next steps:

1. Notes from Ben Gregori

Identity for a Refugee - Survivor identity needs

1. High demand for security and recoverability. Survivors have various devices over their journey (from trafficker to recovery; across multiple types of device) – needs anonymization.

2. Passwords are not good – trauma impedes memory

3. Recovery mechanism (based on Kaliya’s paper) might work by using facial recognition without storage (using a hash – enables recovery with face without storing data; also allows revocation for corrupt staffers or other system participants with access to these credentials)

- overlap with refugee population (no identity and needing to create one – used name has no record, a lost identity – documentation is lost, reborn identity – assuming a new identity to escape trafficking space, guardianship – those who cannot speak for themselves such as minors) - all of these conditions must be considered.

- The role of a certified identifier (by aid staff, camp staff, etc) can provide the verification for identity.

(League of Nations created passports after WWI to reconstruct identities based on their skills, financial, and educational backgrounds) – how do we extend this?

- Trust waterfall strategy: people who already have social trust use your identity wallet, and once they use it, they can help their patients/students to help train/educate
- Can we digitize a personal dossier of medical information so that individuals own medical information instead of doctors/hospital systems (Bangladesh works like this).
- Facebook in Myanmar has filled the place of digital identity for Burmese citizens; these systems were not framed for any consent by the individual and there is little education on it. However, it’s functionally a huge improvement

There is always an infrastructure question when working with vulnerable populations (such as in refugee/migrant situations) – how do we help people where there is no mobile connectivity?

Digital identity frameworks (or imposing one on a population) created by authoritarian governments enable majority populations to persecute minorities. So developing this should be independent of government, but enable government to plug in.

Lost identity workgroup – how do you secure biometrics when they can be forced?

There is a competing priority between future-proofing a identity system and creating a well-thought system that cannot be abused in different contexts in the future, versus helping refugees and vulnerable populations immediately when some populations may be willing to sacrifice privacy for greater benefits, such as receiving needed aid.

Guardianship is really important: mothers may adopt children out of good will that have been orphaned or their parents cannot care for them, and that may be a valid context. However, being lose with guardianship leaves holes for trafficking.

Community attestations – how can we use community attestations (that can localize identity frameworks to the relevant popluations) in conjunction with government/official sources of attestations, with different weights applied to the importance of each attestation relevant to the situation and the level of trust a verifier is willing to accept (eg community attestation may be fine if there are no other alternatives and it’s a low-risk consequence if there are errors, versus more community attestations AND GPS location verification AND a government record to take a child across national boundaries).

ITU: SDG Digital Investmetn Framework – removes the dilemma of solutions either being for-profit or open source with regards to questions about sustainability and adoption.

“Beyond the Hype of ID4D” – short paper to gather critical literature of ID4D

UNHCR virtual summit portal (DIF)

How does this solution look from a tech perspective?

2. Notes from Pam Dingle:

- Jessica -

- Kristina  -

- Jax - and


- Sharing experience for people with marginalized identity.  Working with people who are at risk such as refugees and domestic violence victims (survivors of gender-based violence, human-trafficked)

Annie Cannons trains victims referred from shelters to be developers.  Some are for hire and some are the ideas of the people themselves.  Ideas coming from Shelter victims often cover "how can we improve access to services"

Issues that came up:

- every software solution used at an agency is built for the agency, to help them manage cases and people

--- not designed with the idea that a survivor could use it

- some folks are born into slavery and have no real identity

- some folks are PHds:

People end up at the same facility many times, because they are being re-trafficked.  The system doesn't understand when we are helping the same person vs helping different people. 

Can we put agency into the hands of the survivors? 

Can the system check if you are eligible?

In trafficking scenarios, devices change all the time.  Sometimes traffickers sniff traffic in local areas.  Therefore there is a high bar for multiple devices, device security

Phone may come from  trafficker or by Verizon home line.  Anonymized data is important, can't trust the phone

- if you are highly traumatized you are probably not going to

- eg app loooks like a period trafficker to get past supervision but has extra features.

Need face recognition w/o storage of a face

- sometimes actors at the help orgs are in fact traffickers or abusers

--- need to revoke but also to track WHY they were revoked

Use cases for Human Rights Causes: 

1) No identity - you need to create one

2) Lost identity -- you had identifying docs in origin country but none in new country.  some might be reproduceable

3) Reborn Identity -- a kid who was in forced servitude might need to start new

--- somebody might need to be trusted or not trusted

4) Guardianship for those who can't speak for themselves

eg:  78% of pimps in bay area were themselves abused as children

Earning money means earning food and water yourself - how to create an identity that gets you there?

- gov'ts don't always do this

Trust Waterfall strategy

- people who you trust start to use your wallet (doctors, case workers, teachers)

- they set you up and then teach victims how to use it

- in bangladesh, patient owns their identity

- paper based, they have a dossier

- can that be digitized

- talking to refugees in Greece, Syria, Lebanon

--- each camp is very different

--- eg in Bangladesh, some have smartphones, access to digital stuff

In refugee context, how do you tell what works

- collection is local, doesn't need deep technological learning

- verifiers act centrally to look at the local


- as of 2012, the most a citizen could have is a paper card with information filled out in pen.

- if a kid is caught in a raid, there is no way to get that

- facebook *is* the internet in Myanmar.  Today, facebook is a defacto identity infrastructure for burmese gov't

- those systems have not been framed for consent. 


For refugee:  very often this is an infrastructure question not a technology question. 

- there was infrastructure yesterday maybe, but not today


UNHCR worked with Facebook - if they don't hand over any data, then those people will never have status. 

- if they do hand over information, the gov't can persecute, but Myanmar has a fledgling digital identity effort.

Dilemma: is it better to be known and hated than not known, when it comes to asylum seekers?

If you are fleeing and your fingerprint will expose you as a known identity - what do you do?

Here we talk about user centric identity as privacy preservations but in developing countries, the issue to be resolved is just identity - having one at all.

- eg:  aadhar:  yes, central identity was stored, but purpose was different - goal was to give the ability to billions of people to prove they are a person to the government.  Lots of people in lower classes especially gained huge benefit even if the privacy concerns weren't perfect

-- access to services is critical, not a privilege

Women coming in to these programs have never had privacy in their lives, ever.  They are asking questions like "how do I know that my email is being monitored". 

Note from Joy - Denmark is 50 years into their national id and the privacy issues are coming home to roost.  Need to build it in

Neha - it is important to understand the scalability and remember why this was done.

Jennifer - refugees really do care about privacy because their personal safety depends on it. 

Part of the consequences of any technology we put in place needs to take in the GRAY area

- eg - woman has 5 children, 2 are her own, 3 she claims to protect them.  Some of them might be trafficked.  This gray area is where someone is trying to do good, but some identity requirements might put those children in danger.  We need to be aware of the consequences of the cold beaureaucracy need to be viewed in the lens of protection

Colombia - all the immigration projects right now are tough - colombia is small and refugees are overwhelming.  How can colombians manage the identities for example the children born in the camps.  do refugees care about privacy or belonging to a place (or both)

- in some places, the importance of belonging is huge.  If you have not one single document.

Maybe we end up with an identity solution that can work with a government and one that can work independently - maybe these need to be independently considered

question:  Do we need a global identity for everybody?   

-- some countries are abusers.  We need to

Question: does iris scanning work as well for dark eyes?  Answer yes, but still working on what happens with children under 5 can be reliably identified? 

An attendee worked on Identity for land registries - somebody came up with the idea of weighted attestations - records from NGOs, camps, community attestations, could we take some of those methods

Global identity:

- Could you trust someone that has shown a cycle of bad activity

- entities that aren't a government could offer an identity to people that might be more valuable

GlobalEntry - works in 15 countries, gives privilege to go into/out of airports

-- adds a risk

Project in Papua New Guinea - iris recognition powers an ATM - just talking to village owner, they could distribute the ATM without the gov being involved.  

Counterpoint - you can't exist without cohabitating with the rule of law. 

Paper currently written that points out - not every identity needs the same level of security, but the services that the identity can become a problem - eg, the level of identity needed to get into a camp vs. the level of identity to get government payouts

- need to determine when the next stage of security is needed.  

- security decisionmakers sohuld be able to determine this. 

Tradeoffs between government support and privacy -- how do you give people a choice?

Most consent today is ONE WAY consent.  There needs to be a way to take it back.

George: as engineers, we try to solve edge cases but this is about real people who are in danger.  We need to start looking at the good of the system and weigh it against the possibility of evil.  We are thinking about edge cases and ways that this can be abused - but in our world there is nobody in physical danger. 

- local tribal policy is drastically different across the world.  We (as americans) should be careful to force our idea of privacy across these very different worlds

Most folks here are for-profit companies.  Adapting a for-profit product for non-profit beneficiaries may be incompatible. 

INitiative is being put forward by the ITU -  DIAL and ITU "SDG digital investment framework" argues for interoperable components, identity, payments, messaging are the core components.  The for-profit and not-for-profit

Smart Africa is another digital initiative

Balacz: Europe is building the worlds largest biometric database -- but users can't opt out, and terrorism trumps these rules in reality.  Once gov'ts access the data, how do you put everything back in

Jessica: It is great that for-profit companies are interested but it is the people who are at risk that need to be involved in the design process.

Cannot do this alone:  need a global lens, need lens of a survivor, of Identity, of enterprise, design thinking

New paper - Hype for ID4D 

How does this work with technology -

- this is a great use case articulation -

Iris scanning BSP