CashCast: A podcast from the CALP Network
This occasional podcast from the CALP Network looks in depth at the critical debates in humanitarian cash and voucher assistance (CVA). If you would like to take part or suggest topics for future episodes, get in touch at info@calpnetwork.org. CALP is a global partnership of humanitarian actors engaged in policy, practice and research in CVA.
CashCast: A podcast from the CALP Network
2.6: Will New Technologies Help Deliver More and Better Cash?
Tech brings new opportunities, risks, and dilemmas. Could the combination of technology and cash bring transformative change and finally make aid more people-centered? What are the risks and what are the opportunities?
Host: Karen Peachey
Guests: Laura Walker and Chris Hoffman
Karen Peachey: Hello, and welcome to CashCast. I'm Karen Peachey. In this series, we've been asking what's stopping cash from transforming humanitarian aid. In this episode, we will be looking at cash and technology. Much like cash, technology is a force for change in humanitarian action. New technologies bring new opportunities, new risks, and dilemmas with big questions around ethics, expertise, efficiencies, and more.
With the use of artificial intelligence, AI, rapidly gaining momentum in many walks of life, we can assume it will create change in cash and voucher assistance as well. Whether you find new developments exciting or concerning, technologies are changing how we do things. Could the combination of cash and tech bring transformative change and finally make aid more people centered?
Or is it all hype? In either case, what are the risks? And what are the opportunities? Let's get started.
Discussions about tech in humanitarian action often focus on the way agencies are using technology in their work or to further their own ends, whether that's in the delivery of cash, accessing hard to reach areas or in verifying who people are. But something that's often not considered is that affected people themselves will frequently turn to technology as a means of coping with the crisis they find themselves in.
So let's hear from one of our guests on this.
Laura Walker McDonald: There are more and more people out there looking at how can we acknowledge that people who are affected by natural disasters and conflict are digital actors in their own right. And how can we, you know, leverage some of those technologies to help them in when they need it, or to help them help themselves, how our community is already using this technology.
Karen Peachey: That's Laura Walker McDonald. She's a Senior Advisor, Digital Technology and Data Protection at the Regional Delegation for the United States and Canada of the ICRC. Today, Laura has removed her ICRC hat and is talking to us as someone with great experience and knowledge related to inclusive technology evidence based decision making, ethics, responsible data use, and much more.
Laura Walker McDonald: If you look at Ukraine, for example, where communities are self organizing using social media and messaging apps like Telegram to figure out how to get assistance and migration routes and whether or not and when to return, these kinds of things, or whether you're thinking about volunteer radio stations in rural contexts in East Africa, these are all technologies that sort of spontaneously organizing to use for things that matter to them and people use those technologies to the limit of their possibility and very sort of thoughtful and nuanced and organic and opportunistic ways, and then we come in and we're saying, well, that's, we'll build you a smartphone app.
Well, it may not be the right move. I think we come in with these conceptions. Of what we could do and sometimes they're wildly misaligned with what is practical or possible or the operational realities. Can we really marry the potential of technology in the really intensely unique local way with the, the intense pressure to deliver you know, aid to, to people at, at greater scales and in, in more complex environments than we've ever seen before.
Karen Peachey: What will the humanitarian system need to be doing if it's to succeed at this?
Laura Walker McDonald: We're going to have to get better at plugging into and being aware of therefore local technology realities. And, and we've seen some of that with, uh, even something as simple as messaging apps.
I mean, how many times have we gone into a new response and only to find that the assumption that we can send everybody a message on WhatsApp is wrong? Because people don't use WhatsApp in that particular setting, so, but WhatsApp is the technology that's been approved by head office. So, um, this is a fictional example, but, but it happens all the time.
Karen Peachey: It sounds like currently there are some interesting initiatives going on, but as with many aspects of aid, There is an issue of meeting people where they are, and using the technology that they are using, instead of requiring them to adopt our systems. Chris Hoffman, CEO of Humanity Links, who you might remember from the previous episode, put it simply.
Chris Hoffman: I think one of the biggest issues that we're facing, especially in the inefficiencies, is that we can't meet people where they are. And where people are today, and all the studies, GSMA studies, etc., are finding the same things. People are using their phones. If everybody says I want money over the phone and we still do a traditional cash programming, whose fault is that?
It's on us, right? And so we need to be able to listen better, but also have a business model to operate within that meets people where they are. How do we leverage the phone to register people? How do we leverage the phone to get them cash? How do we leverage the phone to hear about what their experience was to make our work better? How do we use complaints and feedback to through our loops to do better?
Karen Peachey: But it's not all bad. There are many situations where it is working and where humanitarians are successfully integrating with the systems that people are already using.
Laura Walker McDonald: Humanitarian cash using existing national banking systems is a great example.
There is no system more battle hardened than a national banking system. So let's use that. People have their own accounts. They're well incentivized to use it. There's an audit trail. It can be very efficient eventually, once you get it all set up, I say this knowing I'm sure there'll be lots of groans from people who actually implement these programs because of course nothing is straightforward, but using existing systems means that they have the potential to to outlast the response that they're you're not having someone set up a whole new thing or get to grips with a new technology in order to benefit from what you're doing.
Karen Peachey: But often, Chris argues, when it fails, the issues stem from humanitarian organizations using out of date systems and not being aware of the fact that there are better ways of doing this.
Chris Hoffman: What is happening, I think, in the humanitarian sector still today, which is still the traditional way of doing cash, is that an organization negotiates with an FSP, a financial service provider locally, and then sets up a bank account or, you know, there's obviously myriads of things that they do to get cash to somebody that, that I still think is a bit archaic. That's still kind of 20 years ago thought. Now we have payment aggregators that are, that are operating throughout the world. There are many of them that aggregate 450 FSPs over one solution.
Karen Peachey: FSPs, payment aggregators, in case you're not following, here's a quick explainer.
Chris Hoffman: So imagine, um, these, these large internet companies right in the world that, uh, need to make payments to tens of thousands of people all the time, right? So they're having to make constant payments to bank accounts in every country in the world.
So how can one company be able to do that? Can you imagine if we had to do that? Then we would have to go to every FSP and, and negotiate with every financial service provider. So instead of doing that, what aggregators do, they do it on their own because they make money off of it, right? They, they charge a fee, but then they're the ones that hold all the relationships with all the financial service providers.
They're the ones that do all the exchange, the foreign exchange, the FX, um, in any of the countries that we operate in. And so instead of having to negotiate anything you go through a due diligence process and, and then become a person that can engage on this aggregation platform. And then you can distribute plenty to anyone that you want through, through, through their pathway.
So they hold, as I said, the one that we use holds about 450 banks, globally banks, Western Union, Money Gram as well. They also are part of the largest, uh, mobile money provider and mobile pop up provider in the world.
Karen Peachey: So for Chris, the use of aggregators for payments brings efficiencies for organizations, with less contracts to negotiate, and it benefits the person receiving cash, as they can choose which payment provider they want rather than being limited to one chosen by a humanitarian organization, technology has evolved beyond what the humanitarian system is using. So why are these systems not being used? Chris says, in part, it's often due to a lack of awareness.
Chris Hoffman: When you talk to NGOs and you start mentioning these things, they're like, oh, I've never even heard of that. So the reality is, is that the technology has moved much faster because of the business sector understanding. The impact that being able to get people cash quickly can have in non emergency situations, but we have yet to be able to bring that technology in. And so I believe that by creating verified credentials that we can push to financial service providers, adhering to their KYC requirements will allow us to expedite the cash delivery process very, very fast. And why not start to use the tip of the spear technology that's going to grow and continue to grow throughout for throughout the next two decades. To, to be able to start to use it.
Karen Peachey: There are many opportunities to leverage technology. For example, a recent report commissioned by the CALP network outlined how humanitarians can support people on the move, including a wallet based approach where know your customer credentials can be shared with financial service providers, possibly using blockchain. So why not use the tip of the spear technology? If this technology is there and can improve things, should we not implement it?
Chris Hoffman: We like to spread out as much as possible. We like to spread our risk, so to say, right across many different entities. But this idea of efficiency in smaller numbers is not really part of our work.
We want to go through a procurement process in every country that we're in. We want to have a number of people look at it in every country that we're in. But there's something wrong with that. And the thing that's wrong with that on my side is that we all do it with Microsoft. Why, why is it that Microsoft who holds all of our data and all of our emails and everything else we're okay with, but we're not okay with doing it for anything else.
Karen Peachey: Chris raises another interesting point here. While spreading risk across multiple providers may be less efficient. Does this outweigh the dangers? Is it too risky to rely heavily on one supplier? Are humanitarians willing and able to navigate the complex financial and contractual arrangements involved?
Is this even possible using current procurement processes that often hold financial service providers at arm's length? Might it be that we have failed to be critical enough of the technology that our own systems rely on rather than just those that we share with affected people?
Laura Walker McDonald: I did see some analysis that suggested that the pathway that technology is on is towards greater concentration of power in these very large tech firms who are pretty lightly governed just because they are so large, it's hard for any one state to actually govern them.
It's very hard for a startup to invest the kind of money or to have access to the computing power or the data that the big tech firms do so they have an advantage in developing new technologies because they already have an advantage. We all have to send email. We all have to use spreadsheets. And the providers of those technologies have been from a relatively small list and no one's really thought much about it. The same time you're seeing countries and economic blocks trying to regulate in a way that is increasingly fragmented and reflects different value systems and priorities around how technology should work and how it should work in relationship to government and to society as these companies grow in size and their interaction with things like geopolitics becomes more complicated as it's already doing that we have to ask real questions about what that means. And when we're trying to operate in environments that are maybe within a grouping that deals with these technologies in a certain way, so, for example, requires everybody to use non Western technology for some reason, how are we going to operate? Are we even well set up for that?
Karen Peachey: As well as improving payment processing and giving recipients more choice, Chris believes that the use of aggregators can also be used to address other issues.
Chris Hoffman: We also work with a digital identity aggregator as well. So what it is, take it to the humanitarian sector, right? We register a person in need.
Now, when we register that person in need, we verify who that person is. So we issue a credential to that person. And so that means that sitting in the cloud then is this registration of this person. And then that person might have already gone to the gym and then the gym creates a credential for them as well so that they can enter in the gym.
And then the government, where they're living, also issues a government credential for them and a tax credential for them. Now, all these credentials are sitting in silos. So you've registered here, just like in a humanitarian setting, you registered with Oxfam, you registered with WFP, registered with World Vision, right?
So now think about all those silos. Now all those silos nobody can query. So I have no idea what you've told Oxfam. I have no idea what Oxfam's been providing you. But if we can aggregate all of these credentials in a secure way where it's you own it as the person that registered. Um, then I can query Oxfam's information about you, but I can only query what, what I'm allowed to see.
So I can say, has Josephine been assisted by Oxfam and I query your credential and they say, yes, she has. And then I can say, did Josephine receive cash assistance from Oxfam? And then, then says, no, she hasn't. And say, oh, okay, well, actually, then you are able to receive cash assistance from me at World Vision.
And so we, we are able to aggregate all of these identities anonymously. So, I can query the data, not look at it, but I can query it. But when it gets to Josephine, Josephine has to say yes to me. So I, I would scan a QR code, for example, on her phone, or she would provide me a code if she doesn't have a smartphone.
I type that in, I can verify that it's her either through biometrics or her face or any other unique identifier that we, we agree to. And then she gives me permission to query so that she can get another service. And, and so all of her data is hers, she owns it. But then every organization that she's engaged with holds behind them that identity so that we can create, help her, help her even more deeply.
So it's, it's, it's the future. Governments already do it. I'm sure in, you know, the countries that many of us are in, um, there are digital identities either on a card, you know, where you have an NFC card, um, or, or on your phone. Um, like here in the Netherlands where I live. The government has a one distinct digital identity credential that I carry around on my phone and anywhere I go, for any government service, from the doctor, to the pharmacy, to, to the bank, I show them that credential and I'm in. And they know everything, they know that I'm a verified person here.
Karen Peachey: The system that Chris described sounds like a compelling solution that could potentially streamline the experience of interacting with humanitarian organizations for people affected by crisis.
And no doubt it would have benefits for the humanitarian organizations themselves. But is this something that would serve the population at large well? And how does this affect the power imbalance between those offering the system and those using it?
Laura Walker McDonald: If you are providing humanitarian aid to someone and saying, you can have this support if you use this system, is it really consent if there is no other way to get support and you really need support?
So people might be entering these systems reluctantly, and they may even tell you that they don't want to do it. You know, I've read about programs that people have said, I really don't like this, but I'll do it because I need this help that you're offering me. It does then lead you to ask who is excluded by those systems and what are the risks?
And I think having a clear eyed understanding of both of those questions is really important. So as humanitarians, we say we wish to do no harm. So if we are, you know, requiring people to have a bank account that's maybe tied to their legal identity, They may not have a legal ID. They may not have a bank account.
They may be unbanked for good reason. They might be, um, you know, needing to stay under the radar for other reasons. They might be in a population is doing that or a group like LGBTQIA plus people who might have challenges, um, using their legal identity for other reasons.
Karen Peachey: So there may be hurdles that stand in the way of affected people wanting to use these systems. As Chris alluded to the technology, this is built on would need to be highly secure. Otherwise there are some serious risks.
Laura Walker McDonald: Data I often say is a toxic asset. It doesn't age very well. What I worry about a little bit. Or a lot is that we're generating data that may not pose a risk right now, but in the wrong hands or in future years might subsequently pose a risk.
And to go back to the queer community, LGBTQIA plus people are facing a much tougher environment in some places in the world now than they were even a few years ago. And so if you have generated data through some health program or something around people accessing a particular service or, or getting a particular type of help or the family units that they associate with. And so the wrong person gets access to that information and can mine it for information, behavioral information. It's also getting easier to process, again, the role of AI here, getting easier to parse through lots and lots of data to find things out, um, and to do social network analysis. You might be actually painting a target on someone's back.
And I think we've insufficiently thought about that and we're insufficiently humble about the risks we might be actually creating in terms of generating lots of data that then could subsequently be used differently in future by a different type of actor or even the same actor, but with a different intent.
Karen Peachey: So how do we square the circle on this?
Laura Walker McDonald: If we are humanitarians and we want to provide impartial aid to everybody, uh, on the basis of need, and then we do something we know is going to exclude a group. Then we have to ask, is that something we know is going to happen and we have to just be okay with that? Or is that something we're going to try and meet that need some other way?
Karen Peachey: Laura's point raises a wider question here. How do we choose what technology to adopt? When it comes to decision making about technology, questions about efficiency, effectiveness, and costs tend to prevail. But I wonder how often questions of principles and ethics form a critical part of the assessment. And how well prepared are humanitarians to navigate the governance aspects of technology in a sector that is short on hard and soft digital skills?
Laura Walker McDonald: As humanitarians, of course, we have the humanitarian principles, but you may also have other commitments to things like decolonizing aid and your practice there, or intersectionality and diversity and leaving no one behind. These might be things that are really important to your organization. And if that's the case then that should dictate your technology practice. So if you're really concerned about not leaving anyone behind, then you have to think about how technology programs make assumptions about things like gender, and what roles people can play in a household, and what a household looks like, what community leadership looks like, and how you're interacting with those things.
Very basic stuff that, you know, monitoring and evaluation people or research people have been grappling with for a long time. But in technologies, we encode those assumptions in ways that we think are neutral, or we think are we're not doing anything, we're just writing down reality, but actually we're shaping it as well.
Karen Peachey: Laura believes a more robust learning approach is needed to assess different technologies, to understand where and how they can be used appropriately. As part of this, she argues, there is need for good feedback mechanisms to understand how things are working for people on the receiving end, and there is need to involve professionals from different specializations so that technology can be well implemented.
Without all this, her theory is that we might be increasing the risks and the challenges that come along with technology in the humanitarian system.
Laura Walker McDonald: The things I worry about most are where we're jumping into use cases, either not having thoroughly tested or, or grounded ourselves in small pilots first, and then also sort of ignoring what we do know about things like bias in AI, because AI is built on data and data sets are biased. And so making an assumption that you could build a model that say analyzes financial flows that was designed for the US and then use that in a country that has a much bigger informal economy, for example, I think that's a big assumption. So lots of testing is required. However, you know, greater accessibility of AI models to do things like crunch the numbers and improve our logistics supply chain. That's, that has great potential and potentially means that we can do more with less as humanitarians reach more people because we're finding efficiencies in other ways. I think there are real benefits to that. But also, you know, the side would say the same thing. Let's try to understand the technology as much as we can. And understand its interaction with its environment, with the complex and very rapidly shifting realities on the ground for people who are affected by disasters and conflict.
Karen Peachey: In addition to how we decide whether or not to use certain forms of technology, there's also the question of how we regulate its use once it's in place. From Chris's point of view, understanding how data is processed is essential. He feels there is need to put robust governance structures in place so that progress is not optional.
Chris Hoffman: We need to understand how data is processed by the technologies that we engage with. And then be able to put the governance structures, as you said, in there and standardize it across the board.
And if you can't do it, you can't play. I'll give a great example. I love to golf. Okay? Now, in the Netherlands, you have to have a license to golf. And that's because they don't want you just fooling around on the course. We don't want you hitting somebody's house with a ball. We don't want you hitting somebody else with a ball.
And you've got to keep the flow of traffic going on the course. So you have to, you've got to have a license. So maybe we look at licensure of humanitarian organizations, at least at the international level, you have to register when you go to a country, but that registration is, it's just like a business registration, but all the rules that sit under it aren't there.
And how do we, how do we start to formulate that? And do we want it? I can, Chris Hoffman can say, I think it's a great idea, but then you ask the top 10 CEOs of the top 10 organizations what they think. I'm not sure they could answer the same, but I'm not sure that they would.
Karen Peachey: Modern smartphones have only appeared within the last 20 years. And AI, in its current form, has barely been publicly available for a year. Technology moves and is moving incredibly fast. So trying to understand what the long term trends are, and where we might be in the next 20 years, is really hard. And similarly, in the long view, the greater adoption of cash as a form of aid is still relatively in its infancy.
Perhaps it's too soon to expect transformational change across the system. If that's so, What's next for cash when looking at it through a technology lens? Broadly speaking, Chris sees two areas with great development opportunities for the future of humanitarian cash assistance.
Chris Hoffman: I think one is around frequently asked questions and engaging in service, building service delivery and keeping people engaged with us over that period of time.
So one is around communications. How we curate our conversations, how we store the data from the conversations and how we build out more language models to allow us to engage with people differently as we learn from them. So customer service. So I think that's one place where we really need to be.
Within that customer service, I'm going to put the language bucket in there as well, which is translation. So the idea that using AI, when somebody reaches out to us in language A and we don't have any staff that speak language A. What are we going to do? Well, now you don't have to do anything. You can push it directly through the systems and it will translate it into whatever language the staff member that's looking at it is.
So we can really, we reduce this idea of translation services and hiring 10 different staff from 10 different, you know, languages to have on the call center. We don't need that anymore. I mean, we do for some languages, but in general. The second bucket is the data. I mean, organizations are starting to get data analysis, data analytics, data people on their teams.
The AI can really help us with that. So the idea of looking at an anonymized dataset and telling me what I don't know about that anonymized dataset, because right now what we do is we think about what we want to know from the data, and then we put an analytics model on top of it to either validate or invalidate what our assumption was.
And so when we use AI, we don't start with any assumption, right? We're starting with, this is the data set. Please find me the top five trends on female headed households. I believe that those are the low hanging fruit for humanitarian organizations to engage with in the CVA space on AI.
Karen Peachey: Perhaps cash hasn't had the transformative effect that many had hoped for, but it's likely that more change is coming. And Chris believes that change will strongly involve technology and cash.
Chris Hoffman: Localization provides one great opportunity for us and I think getting that right is really changing our whole model and just having to understand that our whole model is going to shift very fast and very deeply. The 40, 000 person NGO is going to be no more in the next two decades.
I think we're going to have much smaller international NGOs because we will have gotten the model right, I hope. to, to be able to engage much more with local organizations because that's the only sustainable solution to this work. And so I think that the opportunity we have right now is completely upending the HR structure and the way that HR is conducted in the humanitarian sector and start hiring for the future, not hiring for the processes of the past.
You have to be ready for a real change, a human resource change, a procurement change, a way of working change. It's not just what your, what your KPIs are. That's not what we're talking about anymore. We're talking about a whole new way of doing work in this industry. And there needs to be people talking about it and working on it and working groups and whatever else we need to have, you know, to start, start nailing this piece down, both from the soft piece, the people piece, right?
Us, but also from the, how the technology and how that's going to work and how we're going to use that. And then bring the two together so that we can then have fit for purpose non profits, fit for purpose humanitarian agencies in the world, delivering cash, delivering assistance at scale as much as possible using digital means to meet people where they are.
Karen Peachey: So there we are. Huge opportunities to make cash and voucher assistance more effective, more people centered, and more efficient. At the same time, alarm bells are ringing about data management, about the use of technologies in ways that don't meet the needs of users, and much more.
What's clear from both Laura and Chris is that there are ways to use technologies to improve aid. But we need to think more about the ethics and the principles as we scope, test, and use them.
We've covered a lot in this series. Our guests, Ahmed, Chris, Christina, Caroline, Innocent, and Laura, have helped us unpack many issues linked to people centered aid, locally led response, and the way the humanitarian system works overall. What's clear is that change is happening. But it's equally clear that more change is needed.
Our guests shone a light on things that need to change with issues related to shifts in power embedded into many conversations. Many issues have been raised and lots of ideas have been shared that would improve the way we deliver cash assistance and improve the humanitarian system overall. Some of the changes needed sound simple and within reach.
It should, for example, be possible to engage with communities more effectively so that. At the very least, people in crisis know what they're entitled to and when. But if it's that simple, why isn't it happening? One thing that is clear from this series of CashCast is that the solutions are not about more checklists and guidance.
We have them. The issues are deeper and perhaps more reflective of deeper changes needed in the humanitarian system. From the discussions we've had, there's need to shift power in the system and shift it to people in crisis so their voice is heard and acted upon. Equally, it seems there is need to change mindsets and incentives that lead us to maintaining the status quo, or not changing things faster, even though the evidence clearly shows that more change is needed.
At the same time, we need to understand the complexity of change. Changing one organization is hard enough. Changing a whole system requires huge collective effort. There is hope. Much as each guest has outlined issues and problems, they've also offered solutions and shone light on possibilities. I for one believe change is possible, but it will require substantial commitment and collective effort.
Thank you for listening. We hope you've found these reflections thought provoking as we continue to work towards more effective cash and voucher assistance. Agree or disagree with the perspectives shared? We'd love to hear from you. While this is the final episode of this series, there's so much more to discuss.
We look forward Cashcast. Until then, goodbye.