ICRC Bet Pivots to ‘Do No Harm’

Challenge bet, “Digitalization as Enabler to Re-Connect Families in Time of War,” has pivoted. But far from dashing hopes, its decision is likely to have an even bigger impact than if it had continued uninterrupted.

Ten months ago, bet Captain, Charlotte Lindsey-Curtet, Director of Digital Transformation and Data at the International Committee of the Red Cross (ICRC), pitched their bet:

“Within a year, the ICRC and EPFL will resolve three cases in which a missing person is identified through analysis of publicly available data, in compliance with safety and data protection standards. Insights gained will provide evidence of the overall potential of data analytics to find missing persons, and identify which sources of information and technologies have the largest potential for identification of missing persons.”

Recently, during an overnight layover in Zurich, Charlotte took time to layout ICRC’s reasons to retrench at this late juncture. She also had some remarkable news that should validate for anyone contemplating a challenge bet – regardless its difficulty or eventual outcome.

Surprising turn of events

Charlotte can tell a good story – as her May 2018 TEDx Selnau talk attests. But when we met on a cold January in a Europaallee bike shop cum coffee bar, she gave me the bullet point version with no warning. “The complexity of what we’re trying to do is really hard. It requires a lot of data and algorithmic techniques that are having to be tested and adapted to the realities that we’re seeing. And we will not achieve it within the bet time frame, but we’re still going to do it.”

Taken aback at her blunt assessment, she proceeded to whiplash my lowered expectations by confiding, “What came out of the bet challenge interestingly, was a funder who said they were ready to fund our challenge.” Charlotte is optimistic of their bet’s future sharing how the donor’s motives aligned with ICRC’s objectives. “They see this as such a concrete way of demonstrating digitalization – of demonstrating the human impact of technology.”

People first

Challenge players would do well to take notice of ICRC’s example. She continued, “The bet was a means for us articulating if we could get interest in something that was humanitarian, the use of technology in a very people focused way. It’s not about technology per se, but about people.” Before diving deeper, she reiterated, “We didn’t pick an easy challenge. Otherwise I don’t think that this would have been worthwhile to do, but we are in it for at least the next three years to see this through.”

Scrambling to revise my prepared questions on the fly, I looked across the table to her colleague. Expecting a data analyst, the young man unsettled my poise further by introducing himself as Yannick Heiniger, Partnerships Manager at ICRC. Correctly interpreting my raised eyebrows as a sign of confusion, he clarified, “As partnership manager, I love to define myself as a matchmaker between needs that we have and the private sector, governments and their ability to come and support. We don’t have the expertise to come up with (all) solutions. We need to partner with others and come up with new ways of responding to key challenges. My role is to match key needs that we have as an organization with what is being offered by different sectors.”

Privacy is paramount

My thinking was pushed further away from innovative tech as Yannick echoed Charlotte’s position why they are pivoting. “Our approach to digitalization, our priority is not to find the technology that would allow us to come up with ‘the three matches,’ he emphasized, “but to ensure any technology is compliant with key data protection requirements with privacy at its core.”

His matchmaking skills are not limited to people, I learned later, after directing me to efforts the ICRC is taking to raise warning flags among other humanitarian organizations if data privacy is not kept paramount. Because the same technology used to alert aid agencies of a new exodus of refugees leaves a wide digital trail which can be accessed and misused by third parties, potentially putting people needing or receiving humanitarian aid at risk.

Do no harm

In a joint report by ICRC and Privacy International, quoting Charlotte, “Collaborating more closely with experts like Privacy International can help us to better mitigate these kinds of risks, in order to do no harm in a changing digital environment.” Her ICRC colleague, Philippe Stoll, Head of Communication Policy and Support added, “Technology is crucial if we want to engage with and better serve the needs of people we can’t physically access, but using these platforms means creating an information trail we neither own nor control, and that’s something we must get better at anticipating.”

As my coffee grew cold, Charlotte explained even though it represents a small percentage of their global activities, “… algorithmic reading is something that we’re working on quite strongly.” Such readings she expects, will improve understanding in the environments where the ICRC operates. Which sounds perfectly feasible, but how exactly? Web crawlers? Google alerts? White hat SEO?

EPFL adds their AI

Pressed to give specifics, she offered, “with EPFL, we started to develop a very specific terminology data set, linked to ICRC: protection, medical, hospital, bomb, attack. All the terms you could think of that were linked on – and develop that into an ontology of language. Then run it through public data sets.” BTW, ontology means “showing the relations between the concepts and categories in a subject area or domain.”

Asked for his assessment as bet partner, Professor Karl Aberer, Director EPFL Distributed Information Systems Laboratory (LSIR) replied, “It is a highly rewarding challenge to put the phenomenal recent advances in artificial intelligence at the service of humanitarian action and to help vulnerable people by better interpreting the available data.”

One last selfie

Still keen to understand how analysis of public-private data may one day help reunite families, Yannick matched me with Charlotte’s Strategic Technology Adviser, Vincent Graf. A mathematician by training (EPFL MSc) and 20+ year ICT specialist, we spoke by phone a few days later. “Data analytics is kind of a broad topic,” he said. “What we’re talking about is to use all kinds of variable data to solve cases of missing people as part of our mandate, which is either personal data voluntarily provided to the ICRC and also data available on public sources.”

He cited the on-going European migrant crisis as an example. “Somebody who boards a dinghy, tries to get across the Mediterranean. The people that board that particularly dinghy have all this at the same place, at the same time.” His words conjured an image of anxious faces on a dark beach, in front of an overcrowded wooden boat; the glow of their phones illuminating hope and fear as they send one last selfie captioned: ‘our boat is leaving!!   I love you so much!!!!’

Vincent elaborated on the migrant scenario, “Mostly likely, they have relatives, because it’s rare that you travel alone, a brother, cousin or friends from the same village. All this kind of data, broadly speaking, is available for making inferences and trying to find out what are the relationships between those missing people and where they come from. Who did they know?”

Connecting the (data) dots

He did not use the word “inference” lightly. Oxford’s definition is “a conclusion reached on the basis of evidence and reasoning.” No small feat even for a court of law where evidence is vetted under oath and reason decided by jury. He summed up ICRC’s approach as, “putting together data sources that we have firsthand – because we get a lot of people coming to us, but also, secondhand inferences, about other things that can help us to more precisely understand the journey these people have come through.”

Curious if their solution was taking form as a platform or what? He answered, “A platform is a word that can mean many things and that’s actually the purpose.” He allowed, “By platform, we mean a system status bespoke to the needs we have here – but also aggregating different types of technologies and tools and rules so that it facilitates the work of our experts.”

Information ethics

Vincent was more definite what the new system would not be: “It’s not going to be a mobile app. It’s not going to be a simple website where people put in a name and search. It’s not going to be a competitor to Facebook where people look at each other.” Because a lot of people work on these mission cases, he continued, “it has to facilitate the work with people who are not technical, not data or computer scientists. And by platform, that’s what we mean. It’s something that is not about just the information or just the main findings. It should be something which aggregates all this information and these tools.”

Asked why the project needed to extend its timeline, Vincent cited three nonnegotiable issues needing enough time: technical, legal and perception (trust). He summed up, “We have to think of the ethics if we use our data sets … instead of just using (any) systems, we realize we need to be involved in the design of these systems.”

Paradigm of trust

Unable to find a migrant or refugee family willing to speak on record, Yannick matched me with Marc Studer from ICRC’s Restoring Family Links (RFL) program. The magnitude of ICRC’s dedication to data protection kept growing as Marc politely rejected my appeal for access to a case family. But he was quick to answer, ‘what can people do to support ICRC’s mission?’ He suggested helping promote Trace the Face, one of many RFL initiatives, “… to reduce the time between persons who lost contact with a relative, to learn about our service and use it to make our service better known to migrants and their families looking for missing relatives.”

No stranger of digital innovation or the pain of not knowing the fate of a missing relative, Marc reflected, “I think one of the challenges is to find the appropriate way we use technology. Because there are many possibilities, but we have to be sure when we use technology that it doesn’t harm the people we want to help.”

An hour later, Charlotte and Yannick moved on to their next rendezvous. The stress of capturing my sources on tape over, I ordered a fresh cappuccino and mused on the emotional rollercoaster their interview had taken. One moment aghast, they were abandoning their challenge — the next grateful, that a trusted organization like ICRC is doubling down on their bet to help us navigate the new paradigm of trust in a digitally-driven world.

The post ICRC Bet Pivots to ‘Do No Harm’ appeared first on digitalswitzerland.

Join Us!

If you would like to get more information about digitalswitzerland or become a member with your organisation and participate at Leap, then get in touch with us!