In this episode, Lisa and Fawad discuss the impact of climate change on global migration trends, how the UN’s International Organization for Migration provides solutions to enable safe, orderly, and regular migration for those affected by climate change, and how the collaboration with Snowflake assists in streamlining this process through the use of data analytics. Lisa and Fawad dive into their partnership, and how they use predictive modeling and evidence-based policies to identify issues early and target assistance where it’s needed most.
In this episode, Lisa and Fawad discuss the impact of climate change on global migration trends, how the UN’s International Organization for Migration provides solutions to enable safe, orderly, and regular migration for those affected by climate change, and how the collaboration with Snowflake assists in streamlining this process through the use of data analytics. Lisa and Fawad dive into their partnership, and how they use predictive modeling and evidence-based policies to identify issues early and target assistance where it’s needed most.
---
Snowflake Data Cloud Summit
When: June 3-6, 2024
Explore the world of data, apps, and AI collaboration at Snowflake Data Cloud Summit in San Francisco from June 3-6, 2024. Hear valuable insights from data and AI experts and business leaders, while discovering the limitless possibilities of data, AI and application collaboration for your organization.
BUILD 2024
Dive deep into the world of apps and generative AI from Snowflake BUILD. Catch up on the latest announcements focused on building apps, data pipelines, and machine learning workflows in the age of LLMs.
Learn more and register for a BUILD event in your region.
[00:00:00] Producer: Hello and welcome to the Data Cloud Podcast. Today's episode features an interview with Lisa Lim Aken, Senior Regional Thematic Specialist for the UN's International Organization for Migration, IOM, and Fawad Qureshi, Leader of Snowflake's Sustainability Initiative, Data for Good. In this episode, Lisa and Fawad discuss the impact of climate change on global migration trends, how the IOM provides solutions to enable safe migration for those affected by climate change, and how the IOM's collaboration with Snowflake assists in streamlining the process through the use of data analytics.
[00:00:38] Producer: So please, enjoy this interview between Lisa Lim Aken, Fawad Qureshi, and your host, Steve Hamm. Explore the world of data, apps, and AI collaboration at Snowflake Data Cloud Summit in San Francisco from June 3rd through the 6th, 2024. Hear valuable insights from data and AI experts and business leaders while discovering the limitless possibilities of data, AI, and application collaboration for your organization.
[00:01:07] Producer: Learn more at snowflake.com/summit.
[00:01:12] Steve Hamm: The United Nations recently concluded its annual climate change conference. This one called COP28, which was held in Dubai. So a lot of us are thinking about sustainability and how the global community will deal with the changes that are coming due to climate change.
[00:01:29] Steve Hamm: Some of those changes are out in the future, but others are here now. One of those immediate challenges is climate induced migration. Over the past decade, more than 200 million people have been displaced by floods, storms, and wildfires. These are very difficult times for displaced people and for the communities welcoming them.
[00:01:52] Steve Hamm: Our guest today Bring expertise and insights to help address these issues. We have Lisa Lim Ah Ken, senior regional thematic specialist on migration and climate change for the UN's International Organization for Migration, and Fawad Qureshi, leader of Snowflake's cross industry sustainability initiative, Data for Good.
[00:02:14] Steve Hamm: Welcome to the podcast. Thank you very much. Great. Let's start with you, Lisa. Please explain what the IOM is and what it does. Also, what are you seeing now in terms of climate induced migration, and how do you expect that to evolve, and how do you See data helping with that. Okay.
[00:02:35] Lisa Lim Ah Ken: Thank you very much for this first question, which is a very big one.
[00:02:39] Lisa Lim Ah Ken: So maybe starting with IOM. IOM is the UN Migration Agency. And what we do is we work to promote safe, orderly, and regular migration by essentially working with governments to realize the promise of migration. Specifically on climate induced migration. I would say kind of the statement as an ag web works to promote safe, orderly, regular migration, working with governments.
[00:03:02] Lisa Lim Ah Ken: This gives you a clue as to how we articulate also the work we do on climate induced migrations. And that is really very solutions oriented. So we work on solutions for people to stay. We work on solutions for them to move. And then we work on solutions for people who are already on the move due to climate induced push factors.
[00:03:21] Lisa Lim Ah Ken: In terms of Migration and climate change as a thematic and how it evolves. I think the most important thing is to note that for IOM, we don't see migration as a problem to be dealt with per se. I mean, migration due to environmental and climatic push and push factors have been going on for. centuries.
[00:03:41] Lisa Lim Ah Ken: And it's not going to stop, but in the context of the, the currently rapidly evolving, you know, anthropogenic climate change, migration is a lot less predictable now. It's a lot less organized. It's a lot less safe and therefore it's a lot more urgent. And I would say, this is why we need data. Data helps us to understand like the unique contexts that force particular communities to migrate.
[00:04:04] Lisa Lim Ah Ken: And it's a complex relationship, right? Because people's decisions to migrate are influenced by a whole host of factors at different levels, micro levels, meso, macro levels, and et cetera. And data helps us to be more accurate in our interventions. So for example, we develop evidence based policies and programs for anticipatory action and to target assistance where it's most needed.
[00:04:27] Lisa Lim Ah Ken: So basically data helps us to prepare, to plan, and to receive. IOM has a global data institute, and we work, for example, to enhance data for foresight, like developing climate modeling, you know, to identify the level of human exposure to climate hazards, and to identify areas where people might be at risk of climate related displacements in the future.
[00:04:49] Lisa Lim Ah Ken: I hope this answers your question.
[00:04:50] Steve Hamm: No, that's very good. That's very good. And Fawad, please tell us about Snowflakes Data for Good initiative.
[00:04:57] Fawad Qureshi: Thanks, Steve. In our world today, we are constantly seeing the importance of data across all industries and all sectors are benefiting from the use of data. You know, people are talking about data is the new oil, data is water, data is acid, data is as product.
[00:05:13] Fawad Qureshi: All of those, those terms keep on appearing and we keep on seeing that different commercial organizations are improving their businesses, their bottom lines and, and customer experience through the use of data. And what we are trying to do with this data for good initiative is, could we use the same data to help the society and people in general improve their lives?
[00:05:37] Fawad Qureshi: You know, so this is where We started working together with, with IOM, we, we thought about, okay, we have different kind of providers inside the Snowflake Data Cloud with regards to the, the, the weather data. We have got climate change data. We have got other footfall movement data. Could we bring that data together to start bringing different, different insights, insights, which could be used to predict and forecast those people on the move and, and help them while, while they are moving to a, to a new place.
[00:06:07] Steve Hamm: Yeah, so one of the main offers that Data for Good brings is this whole set of third party data that could be, that could be a wide range of data about everything from climate to, to business, to economies. All these kinds of things are brought together in one place. That's the idea. Absolutely.
[00:06:27] Fawad Qureshi: You know, so, so Snowflake becomes that kind of the data broker of those different data sets.
[00:06:32] Fawad Qureshi: Yeah. Bringing that together in a single platform and then offering different kind of capabilities to exploit those data sets in such a manner that could be used for the benefit of humankind. Yeah.
[00:06:44] Steve Hamm: No, but get a little more specific. The data for good initiative. I mean, You're dealing with non profits or what kind of organizations are you dealing with and how do you approach them and kind of what's the deal?
[00:06:57] Fawad Qureshi: Right, the data is not just coming from non profits. The data is coming from commercial organizations as well, right? So, I mean, I'll give you an example. When you talk about in the commercial sector, So the telecom organizations are, are working with retails and, and other markets to provide them footfall data on, on who's walking by a coffee store so that I can offer them a free coffee, some loyalty points and, and whatnot.
[00:07:21] Fawad Qureshi: The same telecom footfall data could be used to predict and forecast how people are moving through a country. To move from point A to point B, so it doesn't necessarily have to be a nonprofit data. The same commercial data set when combined with other useful information coming in from UN, IOM, other, other public bodies could bring in net new insights that were not possible before.
[00:07:45] Steve Hamm: No, I get that. I get that. But specifically, I'm also asking, you're dealing with nonprofits in the Data for Good initiative. How do you deal with them? How do you, what's the relationship between Snowflake and the nonprofits?
[00:07:59] Fawad Qureshi: So, when we are working with nonprofits, we tend to conduct a lot of Data for Good hackathons.
[00:08:04] Fawad Qureshi: Where we bring together different academics, different industry experts, students, and different public bodies together under the same room. Let's try to solve these problems together. Let's bring together different creative ideas from a diverse range of people. And see how we can combine all of these data in a unique way to solve this humanity crisis.
[00:08:28] Steve Hamm: Yeah, I gotcha. I gotcha. By the way, I'm hearing some birds and I know that these birds are chirping outside of a window in Nairobi, Kenya. So that's kind of, that's kind of nice, isn't it? So, all right. Hey, let's rise up. I mean, we're talking at a very high level, but let's go even higher. I'd like to hear from both of you why you think data, data management, data analytics are so critical for dealing with sustainability issues.
[00:08:55] Steve Hamm: What is it about these, these issues that are so challenging and that data analytics is really essential to deal?
[00:09:03] Lisa Lim Ah Ken: Yeah, perhaps I can, I can start us off, um, with trying to answer that question. For me, honestly, the, the complexity of the relationship between climate change and human mobility. It's really very challenging and it's constantly changing.
[00:09:20] Lisa Lim Ah Ken: And I say it's changing because as you rightly said, COP is, COP28 is ongoing now. And I came to COP last week and whilst there's been some, some great progress on, on the topic of loss and damage, I've also seen that Progress towards agreement between parties continues to be excruciatingly slow on, on some very important topics of work like adaptation, like mitigation and food security.
[00:09:47] Lisa Lim Ah Ken: Um, so for me, it's really clear that we need to forge on ahead with developing, you know, concrete solutions and concrete solutions that, that don't just band aid the problems, but are lasting, you know, and help people move towards opportunities and towards self reliance. Reliance, even amidst the continuing, you know, evolving climate change.
[00:10:07] Lisa Lim Ah Ken: And we know that climate change is not going to disappear tomorrow. I mean, it will remain, it will remain for decades, even if parties, you know, miraculously agree and start cooperating towards the goals of the Paris agreement. The changes that we're seeing are really long term. So we also know that all solutions have got to be evidence based to be sustainable.
[00:10:29] Lisa Lim Ah Ken: And without data, you know, without managing and analyzing the data accurately and effectively, we just can't understand the way in which climate change is impacting on people and the ways in which they're moving. So for me, it's pretty clear that that data and everything we do, you know, to collect it, to manage it, to analyze it, is critical in developing sustainable solutions.
[00:10:51] Steve Hamm: Yeah. Yeah. Yeah. And Fawad, I mean, you're, my understanding is that one of your roles in Snowflake is to kind of deal with these sustainability issues across industries, across, you know, different, different clients, all this kind of stuff, across different kinds of data. What's, why is sustainability, why is, why is environmental sustainability, climate change, why are these so challenging and why is data so important to dealing with that?
[00:11:18] Fawad Qureshi: I think one of the fundamental reasons is that sustainability requires a holistic view. of the entire world. So you're not just talking about a single organization or single public body in its, in its isolation. You need to think of the total impact of that organization on the planet and how does the, the planetary variables impact, impact the business.
[00:11:44] Fawad Qureshi: In the world of sustainability, there's a, there's a core requirement called the scope three mission, where you want to understand what is the, what is the footprint of your entire value chain. And that makes it a very, very challenging aspect from a data perspective. I was talking to a major retailer of ours a few, few months ago, and they said, what if the supplier of my supplier of my supplier of my supplier, seven chains down the link is using anti environmental practices under the scope three regulations, I'm responsible for them.
[00:12:16] Fawad Qureshi: Because the principle of sustainability is if you are benefiting from any action directly or indirectly, you are responsible for it. And you can't just say that, you know, I wasn't aware of it. It's, it's, uh, it's happening a couple of continents over. I don't care. That is not, not a viable excuse. And for most of the technology solutions that are available today, you know, the traditional ERP systems, they only have information about your immediate suppliers, the tier one suppliers.
[00:12:46] Fawad Qureshi: If I have a relationship with you, I know what you do. But I have no idea about that tier 7 supplier and that makes it very hard from a data perspective. So, um, so that kind of data brokerage that I was talking about earlier becomes quite important in collaborating between multiple bodies together to build that holistic view because I want to see what is my total impact on the entire planet when I'm conducting my business.
[00:13:15] Fawad Qureshi: Yeah. From the entire value chain perspective. Yeah.
[00:13:18] Steve Hamm: Yeah. I get that. That sounds really good. Yeah. So Lisa, let's get into some of the details of how Snowflake and IOM are working together. Tell us how this partnership got started and, and also about these recent hackathons in London and Nairobi. You know, what's, what's your objective and how does it contribute to addressing climate mobility?
[00:13:38] Lisa Lim Ah Ken: Sure. So I believe the, the partnership Started back in January this year when some members of IOM from HQ, including some of the members from the information and communications technology team, traveled to Silicon Valley and they went for an innovation immersion trip and Snowflake was among the companies that.
[00:14:00] Lisa Lim Ah Ken: And I think they just, we just kind of stayed in touch and started discussing how we, how we collaborate, how we could collaborate, private sector collaboration within the UN. It is still, you know, you know, it's still relatively new, let's say. And, and so, you know, there, there is still kind of a lot of discussion around, you know, what, what we do together, what we can do together, given the very different ways in which we work.
[00:14:24] Lisa Lim Ah Ken: So, yeah. I think over the discussions, the, the idea of the hackathon emerged as maybe one concrete thing to work on, but which can also kind of allow each other to understand, you know, how, how we work. And one of the topics of interest from both IRAM and Snowflake was climate change. Also, because as the discussions continued, we were starting to, you know, get closer to COP 28 and there were discussions about.
[00:14:51] Lisa Lim Ah Ken: About Cop 28. So I was approached as the regional thematic specialist on migration climate change in the Hall of Africa to see if a hackathon on this topic could be developed. Quite honestly, I'd never heard of a hackathon before, but it, it sounded very, it sounded really intriguing. Um, so, you know, I was like, okay, you know, I'm up for this.
[00:15:11] Lisa Lim Ah Ken: Let, let's try it. So it, it started really with kind of investigating the topic, the topic of climate mobility, and as, and as I mentioned, it is. It is really complex. And then, but rather than needing to be addressed, it really needs to be, to be leveraged as there are, you know, lots of opportunities to use mobility and mobility in the context of climate change to contribute towards sustainable development.
[00:15:34] Lisa Lim Ah Ken: So, so the objective, I think, of the hackathon in the end was. was less about addressing the challenges and more about nurturing creative solutions, you know, to support communities that are affected by, by climate and environmental induced migration in the region. And then this, and this was done through multidisciplinary teams.
[00:15:53] Lisa Lim Ah Ken: Who were, who were guided by, by some questions and challenges, and I'll give you an example of a couple of them. For example, how can we use historical data and climate models to predict migration patterns that are triggered by environmental changes? You know, and this would. Help us plan and prepare both migrants and receiving locations and communities, right?
[00:16:16] Lisa Lim Ah Ken: And then another question was something like how can we develop innovative tools and models that can support like advocacy efforts and that can help governments and organizations like to formulate policies which allow migrant communities to contribute to sustainable development, you know, in their host countries.
[00:16:34] Lisa Lim Ah Ken: So these are kind of, they're, they're quite complex questions, but they are the sorts of questions that we continue to struggle with in my, in my thematic. So these were the questions we put down and thought, okay, let's, let's go for this, for this challenge. I mean, of course we, we, we didn't manage to answer all the interesting questions that came up, but, but what was great is that there was a lot of innovative thinking and then a lot of new solutions that we had also, you know, not really considered before.
[00:17:01] Steve Hamm: No, I think that's fascinating. The, the way you phrased that. You know, how can, you know, migrant populations actually help in the countries where they land, if I understand correctly, and that's, and that is a really interesting way to think. I mean, in the United States, for instance, we need workers, we need young workers and where are young workers?
[00:17:23] Steve Hamm: Well, young workers are coming here from. Places that are stressed. So how to turn something that some see as a negative into a positive. The potential is there. It just takes some creativity and, and some data analytics, it looks like.
[00:17:38] Lisa Lim Ah Ken: Yes, absolutely. And, and there is no better example than the host of COP28.
[00:17:43] Lisa Lim Ah Ken: Look at Dubai. Dubai is, is, you know, a migration, a labor migration success story. We can say, I mean, yes, there are a lot of issues with it, but they have done very well on migrant labor.
[00:17:56] Steve Hamm: Yeah. Yeah. That's interesting. And I saw a recent article about how they're changing the way they're architecting buildings there.
[00:18:03] Steve Hamm: With less glass and steel and more kind of absorptive and, and insulative materials and stuff like that. So, yeah, interesting. Hey, so let's go, let's drill into the, the Snowflake and IOM relationship now. So Lisa first and then Fawad, what do Snowflake and IOM hope to accomplish together?
[00:18:25] Lisa Lim Ah Ken: Yeah, so this hackathon was, I think, the very first step in, in two really different entities coming together to firstly sort of understand what each other does, as I mentioned before, and then to explore what we can actually do together.
[00:18:40] Lisa Lim Ah Ken: Both Snowflake and IOM have A lot of sources of data. I mean, IOM collects large amounts of data on, on human mobility, but, but, you know, we store them, we collect them, we store them and we analyze them for probably very different reasons. So I think from the IOM side, it was initially a little bit difficult for us to see how our data can be combined, you know, to develop solutions.
[00:19:03] Lisa Lim Ah Ken: And I think this first hackathon really taught us a lot. About the capabilities of each entity of Snowflake and IOM and we learned that, you know, data can come from a very wide range of sources and can be combined in a complementary way to like to corroborate each other and to develop maybe a clearer picture of the problem statement that we're trying to answer.
[00:19:25] Lisa Lim Ah Ken: So, I mean, I personally also learned that the painful process of like interrogating data can be made much faster with AI and the technologies that Snowflake can offer, and this itself significantly increases our efficiency. It allows us to analyze a larger amount of data and to deal with a huge multitude of different and complex problem statements.
[00:19:49] Lisa Lim Ah Ken: So for me, specifically on the work I do in IOM, as the, the Region of Atlantic Specialist on Migration and Climate Change, I would like us to first of all, I guess, further the further develop the solutions that were already initiated in the hackathon. And then what I would really like to do is to start building the capacities of governments to use them, you know, because technology transfer is, is such a, an important area of work.
[00:20:16] Lisa Lim Ah Ken: End. And I would really like to see, you know, us supporting the transfer technologies and capacities to those that most need them and can, you know, and need to use them the most. Um, so I would really like to see the IOM and Snowflake kind of collaboration reaching the communities and the governments that we support.
[00:20:39] Steve Hamm: Well, that's interesting. I hadn't thought of that. So A lot of the data that you collect is from individual governments, but it's circular. Those governments can also get into the data that you've collected? Is that what you're saying?
[00:20:51] Lisa Lim Ah Ken: To some degree. I mean, all the data that we collect is for specific purposes.
[00:20:56] Lisa Lim Ah Ken: We might collect data on people being displaced in order to, to allow other UN agencies to plan their response. You know, so we would collect data of people entering a particular, a particular area, and then we would, and our data is always disaggregated. So the World Food Programme, for example, would know how many children need specific types of food, or how many women there are, how many pregnant women or breastfeeding women there are, you know, how many men there are.
[00:21:24] Lisa Lim Ah Ken: And so all the data that we collect. It's really for, for response, for something very specific. The data we collect is, is anonymized and we share numbers, but not more than that, with governments and, you know, with other partners, as I say, specifically for, for support reasons.
[00:21:43] Steve Hamm: Interesting. Very, very good. So Fawad, you're, I think you're down into more of the details of actually how IOM is, is going to use the data cloud.
[00:21:52] Steve Hamm: Talk about that. What's, what's the use case and what benefits will the organization get?
[00:21:57] Fawad Qureshi: So, IOF has a very comprehensive set of tools, components, and methodology, and together they call this, this kind of overarching framework, Displacement Tracking Matrix. This has been in operation since 2004, nearly, nearly 20 years now, and they have one of the most comprehensive information about displacement of people across the world.
[00:22:19] Fawad Qureshi: And they are, it comprises of different kinds of survey information, different kind of state of the country and all. And there's a lot of meta information about, about the immigrants and the, and their origins. And destinations in, in that system. So in our partnership, what we said was, okay, how could we overlay some of that commercial, non profit, other type of data sets that we have got in the marketplace to combine it with, with the displacement tracking metrics data.
[00:22:50] Fawad Qureshi: To come up with a solution which could act as an early warning system so that we could predict and forecast that this area is more sensitive. There are, this needs more local support so that we don't want to create chaos in the, in that kind of, in that territory that the datasets that we brought in into this hackathons were some of the telecom footfall data, some of this.
[00:23:16] Fawad Qureshi: Socioeconomic data, some alternate credit scoring mechanisms while, because we are working in the Eastern region of Africa, there's not much systematic banking system available, but there are mobile banking solutions available where people are using a banking transaction over the phone. Could we start using those micro banking transactions as a way of building an alternate credit scoring mechanism, which could then be used as a proxy for economic activity.
[00:23:45] Fawad Qureshi: So if we could start tracking that economic activity at different kinds of administrative levels, admin one, admin two, admin three, we could start understanding, okay, this admin three needs a little bit of attention because the economic activity over the last six months It's slowing down and that could be a sign of something to look out for.
[00:24:06] Steve Hamm: So it's like an early warning system of economic distress. Absolutely. Yeah, that would be really a cool application. So Lisa, you talked a little bit more before about anonymizing data, the importance of that. And I think, you know, my sense is that your organization does a lot of surveying of people who are living in some kind of distress in their own countries or on the move, where trust of, you know, how will this information be used could be a really important issue.
[00:24:38] Steve Hamm: So how do you deal with privacy? In your work, where I think it must be, you know, privacy is a big issue globally, but I would think it'd be even more sensitive dealing with these kinds of migration, people who are migrating or, or, or might migrate. Um,
[00:24:57] Lisa Lim Ah Ken: Siva, I'll talk a little bit generally. about this question based on years of experience of, of working with data collected from really very vulnerable communities and, and following very strict protection guidelines, which, which IOM has.
[00:25:16] Lisa Lim Ah Ken: So I think for us, we are careful to be able to demonstrate our process. of data collection, management, you know, storing, analysis, sharing, and at all stages. We are careful to be able to demonstrate how we follow IOM's data protection guidelines, which as I say, have served us for a very long time. We're very careful in the accuracy of our data.
[00:25:43] Lisa Lim Ah Ken: We check it through, through multiple layers of quality control, and we don't make quantum leaps in, in conclusions or when we make, you know, causal claims. We do our best to corroborate our data across multiple sources wherever we can. And we use a lot of qualitative data as well. Qualitative questioning, you know, to bring context to the data because for us, um, the story is often more important than the data itself.
[00:26:08] Lisa Lim Ah Ken: Or at least, you know, one cannot exist without the other. And, and as a un agency, our priorities always to use data to, to respond to the needs of people, you know, and the governments that we serve. So we're careful to. To really express that there is no hidden agenda when it comes to data with IoWare, we put, you know, our disclaimers up front.
[00:26:27] Lisa Lim Ah Ken: We explain the holes in our data, for example, where we have perhaps less confidence in it. And essentially, it's just about being very transparent about our processes. And these are processes that are standardized and have been tried and tested for decades, really.
[00:26:43] Steve Hamm: Yeah. Hey, so Falwood, I think it would be really interesting if you could talk about an issue that's been out there for years, which is, oh, is data, is it more vulnerable in the cloud?
[00:26:56] Steve Hamm: Because, you know, a lot of organizations handing over their data to another organization is, is, is a very, is very fraught. So talk about that issue. That issue has been out there for a while, but It seems like it's been addressed, but, but how has it been addressed?
[00:27:12] Fawad Qureshi: So, so Snowflake has a, has a principle of what we call as the zero trust privacy.
[00:27:19] Fawad Qureshi: Meaning you trust absolutely nobody in the cloud. And you basically, you build a system in such a way that only access to the data is given on a need to know basis. By default, there is no access to the data and any of the data that goes into the cloud is encrypted both at rest and while it is in motion as well.
[00:27:44] Fawad Qureshi: And those encryption keys could be controlled by people who are not inside the cloud. So like, you know, like IOM themselves, so IOM could encrypt the data that is getting the data that they are loading into the Snowflake platform. They could encrypt it using their own keys and without those keys. That data is just garbage.
[00:28:06] Fawad Qureshi: So you need to have access to those encryption keys that are managed outside of the cloud platform to be able to fully decrypt and see the data in clear. So that, those different kinds of role based access controls, security and privacy controls and access and control mechanisms prevent any kind of data privacy concerns and you can easily build a very robust and resilient solution in the cloud.
[00:28:33] Steve Hamm: Hey, I want to drill down a little bit more on this with you Fawad, just the, you know, we're talking, usually we, we on the podcast talk about Snowflake's relationships with customers. Commercial businesses. Today, change of pace, we're talking about dealing with a UN agency, but, you know, there are governments, UN agencies, nonprofits have different kind of rules and regulations around them and different goals.
[00:29:00] Steve Hamm: So what's the, when you, when you take something like the capability of the data cloud and you start to offer it to this other set of organizations with different. Governance, different rules, financial rules, things like that. How is that, you know, how is it different? And how do you deal with that?
[00:29:19] Fawad Qureshi: There are differences and then there are similarities between both the type of organizations.
[00:29:25] Fawad Qureshi: I'll talk about something that happened just recently. The EU recently passed the AI Act. Right. The idea is to govern all the different kinds of data analytics and AI practices that are running across. Different kind of public and private sector organizations to protect the interest of the consumers and the citizens at large.
[00:29:45] Fawad Qureshi: And one of the fundamental concepts discussed in that ethical AI framework was transparency. So whatever you are doing with data and AI, you should be transparent about it. And the consumers and citizens have the right to get access to the data. They have right to explanation on what is happening inside the platforms.
[00:30:08] Fawad Qureshi: So, when you talk about that kind of transparency and building complete lineage, complete traceability, if I get an offer from somebody, some company, and I go back to that company and say, how did you think I was a good fit for this offer? So that company by law. In, at least in Europe, it's supposed to explain to me what was the process to get to this point where I received an SMS or some other offer through another channel.
[00:30:37] Fawad Qureshi: And that kind of explanation requires a complete traceability in the data. So inside the, the, the data cloud platform, we have complete Framework availability, we have got, we talked about security and governance and all the kind of different kinds of access controls. We have auditability, we have transparency, lineage, all of those capabilities are required, which enable these different public and private sector organization to deliver on their use cases while staying compliant with all the different regulation across the world.
[00:31:10] Fawad Qureshi: When you talk about the public sector organization on how they are different as opposed to a commercial organization. Some, sometimes when you're working with these organizations, they have different kind of geo residency requirements. So they, they're only able to put the data into certain clouds. So many of the cloud service providers have these special additions called the GovClouds.
[00:31:31] Fawad Qureshi: So which are kind of air gapped from the rest of the cloud deployments because they have a different residency requirement. The data cannot leave this physical boundary versus that physical boundary. So those kinds of capabilities are also part of the Snowflake platform. to provide that extra level of security and safety to those public sector organizations.
[00:31:52] Steve Hamm: So it's interesting. So, you know, Snowflake obviously does and wants to do a lot of business with governments. It's like a lot of the capabilities that you put in place for governments are also the kinds of things that these nonprofits or the UN agencies would want to see. So it's not like it's a great leap into this other world.
[00:32:11] Steve Hamm: It's kind of like you've already made the transition. Absolutely. Am I understanding that correctly?
[00:32:16] Fawad Qureshi: It's, it's the, it's a concept of privacy by design. And that's when you build it up from the ground up, that it is independent, modular, and you know, everything is explainable, and there is no, there are no black boxes in the system.
[00:32:31] Fawad Qureshi: That's how you are able to comply to all the different kind of requirements.
[00:32:34] Steve Hamm: Got it. Now, for both of you, and Lisa, you mentioned AI a couple of minutes ago, but you know, here it is. It was just over a year ago that really, The idea and the reality of large language models and generative AI first kind of leapt into the public consciousness.
[00:32:53] Steve Hamm: I mean, it's really been just one year, but what a crazy year it's been. And I know that there's been a ton of work at Snowflake on figuring out, well, how do we take advantage of these new capabilities? How do we integrate with them? What do we put into our, our programs, into our cloud? So if you could talk about how this kind of has changed the world, what new capabilities you're seeing, and particularly what can agencies like IOM, how could they take advantage of these new capabilities?
[00:33:25] Lisa Lim Ah Ken: All right. Thanks, Steve. Well, first of all, I think I need to be really. Frank and honest, the IOM is only starting to navigate the modern world and this is thanks to the direction and guidance of our new Director General Amy Poe. We have honestly been lagging behind and that makes it a very steep learning curve for us.
[00:33:46] Lisa Lim Ah Ken: So I mean, we're excited, we're enthusiastic, we're curious. But we are still somewhat ignorant on the capabilities of large language models and generative AI. And the hackathon was really the eye opener for us. And I think that was the first and probably the most important step and the right direction for us.
[00:34:06] Lisa Lim Ah Ken: So we still have a lot to learn, we have a lot to understand, and we cannot do that without engaging companies like Snowflake. And I have to, and I have to thank Snowflake because they have been so gentle and so patient with us. Yeah. And it's really been. Such a great collaboration. So I think as we continue talking and exploring, I'm sure there'll be, there'll be opportunities for us to take advantage of AI and other technology for the many specializations and areas of work that IOM engages on well beyond the climate
[00:34:37] Lisa Lim Ah Ken: space, right? So, um, I think we just need to keep talking.
[00:34:40] Steve Hamm: Yeah. Interesting. You know, and, and Fawad, correct me if I'm wrong, but Lisa, you mentioned before that when you do surveys of people, human beings out in the field, that you ask them, you don't just collect numbers. You also collect their stories, that kind of thing, that kind of contextual, textual information.
[00:35:00] Steve Hamm: Can be, AI can be used to really kind of do sentiment analysis or, you know, pull all sorts of signals out of, uh, you know, a large array of answers. So Fawad, if you would, and I hope I'm right about this, but if you would talk to us about how AI might be used in, in that kind of situation.
[00:35:20] Fawad Qureshi: If I were to classify the, the human migration in two different categories, there are many ways of classifying this information.
[00:35:29] Fawad Qureshi: If, if I think of, in terms of timing, so I think of sudden migration and gradual migration. So sudden migration is which happens because of a point in time, major event, suddenly a conflict breaks out in certain part of the world, or there's a major category five hurricane or a major flooding and people have to leave in high numbers.
[00:35:51] Fawad Qureshi: So while, while sudden migration has a huge impact on human life and you know, there's a lot of negative consequences of it. From a data analytics perspective, it is quite simple because you know exactly what, what caused it. What was the trigger? What we are trying to solve for here is how can we use then analytics to predict that gradual migration?
[00:36:14] Fawad Qureshi: Because somebody wants to decide to move. They do not decide to move on the day of the move actually itself. It's the events six to nine months, maybe 12, 24 months behind what was happening. Like we were talking about that socioeconomic activity, what's happening in the, in that part of the world. And that data could easily be brought in into an AI model to start predicting and forecasting what is.
[00:36:43] Fawad Qureshi: expected to be happening in this part of the world. And like you mentioned, you know, the different kinds of the survey information that we have in the displacement tracking matrix, and there are different kinds of data sets like flow mobility and all different survey information that could easily be given to different kinds of sentiment analysis, different kinds of tools to understand what is the, you know, what is the why behind the why?
[00:37:06] Fawad Qureshi: What is the real reason? Because there is certain things that those, those people are going to tell you, but then you, what you want to do is what is the systematic Happening at a, at a macro level that is causing these mass displacement.
[00:37:22] Steve Hamm: This has been a, a, a great discussion, you know, forced migration, a very important and complex phenomenon with many causes and many geopolitical factors involved.
[00:37:34] Steve Hamm: So let's, let's kind of recap a little better or, or probe on this a little bit. So both of you, how does data analytics, or how could it help with this big. You know, this big complex issue. So
[00:37:49] Lisa Lim Ah Ken: Steve, maybe I can, I can touch on this by again, how should I say, explaining this, this complexity of climate change and mobility.
[00:37:59] Lisa Lim Ah Ken: Now I will use the term that you just used, which is kind of, you know, a forced migration as a complex phenomenon. And to kind of. To illustrate how complex this, this topic is, I'll ask the question, how do we know when migration is forced? So in the case of a sudden onset hazard, like a flood or a landslide, I think it's, it's obvious, right?
[00:38:21] Lisa Lim Ah Ken: That people are forced to move out of harm's way. That's, that's a pretty easy one, but in the event of a slower creeping onset of drought, for example, where if Rains have failed, where they last longer, the failed rains last longer, where they come more often than they used to. Such as, for example, in the case of the Horn of Africa over the last two years, where we've had a very long and intense drought.
[00:38:44] Lisa Lim Ah Ken: Say a farmer has a failed harvest one year, and the next year he uses his savings to buy seeds. But that again fails. And then the third year he considers his chances. You know, should he try one more time with the last of his savings? Or should he move to the nearest city to look for work? Say he decides to move.
[00:39:01] Lisa Lim Ah Ken: Would you say he was forced to migrate? Or was it a choice? Right? And was the fact that he had children at home for whom he had to pay? Paid school fees, for example, was that part of his decision? Or was the fact that there was a new construction boom in the nearest city part of his decision? So migration is so, so context specific.
[00:39:19] Lisa Lim Ah Ken: I mean, this, this farmer's next door neighbor who might be facing exactly the same problem. might decide not to move because his personal circumstances are a bit different or maybe just his perception of the destination is not as positive as positive. And the question you ask is like, can data analytics help?
[00:39:37] Lisa Lim Ah Ken: And yes, it can to a certain degree. Yes. It can help us identify all the different push and pull factors. It can help us analyze. The more likely balance of factors that could push someone over the tipping point and result in a decision to move. In the academic world, climate induced migration is, is known as a wicked problem.
[00:39:58] Lisa Lim Ah Ken: And that means it's, it's highly complex and it cannot be addressed by, you know, a single entity. So I think the beauty of, of AI tools and, and. Data analytics is that they're able to synthesize really large amounts of information from like a huge variety of different sources and different disciplines so that we get closer to understanding the likelihood of migration, for example.
[00:40:22] Lisa Lim Ah Ken: But for us also, this information in itself is a little bit useless. I mean, what is golden is what we do with this information, right? So we can help governments to ensure that their policies, for example, support people who want to stay. Because we know that there are, uh, there is a great scope for adaptation, so we can support them by ensuring that they inject funds into programs.
[00:40:46] Lisa Lim Ah Ken: for adaptation, or we can help governments ensure that their policies help people to move safely and regularly towards opportunities, like by supporting labor mobility, for example, by working on labor, bilateral labor agreements with neighboring countries, you know, building in robust human rights conditions.
[00:41:03] Lisa Lim Ah Ken: So, so data analytics helps us to develop like more targeted solutions. And the other thing I think which is really great is that Another trait of these, of these wicked problems is that they're emotive, they can be very sensitive, very controversial, and, you know, I'm hearing that AI can analyze to a degree emotive conditions or, you know, topics, but What I really liked, and I, and I learned from this hackathon was that in developing these tools, we're actually forced together in these large multidisciplinary groups, and we're forced together to talk, to brainstorm, to innovate.
[00:41:42] Lisa Lim Ah Ken: So for me, the winning combo was, was both the technology and the collaborative, you know, human centered approach to problem solving, you know, really forcing people to work across their silos and to come together and be innovative.
[00:41:59] Steve Hamm: No, I think that's a great answer. You know, when I think back on the 20th century, one of the, one of the huge trends was this, the building of domain expertise, which brought tremendous advances, but also silos of knowledge and silos of data.
[00:42:16] Steve Hamm: And I think you're right that, that if you bring together, if you say, let's work together on this problem and it's a data project, you bring together the people and the data, not just. The data. So that's really, that's very cool.
[00:42:29] Producer: Dive deep into the world of apps and generative AI from Snowflake. Build catch up on the latest announcements, focused on building apps, data pipelines, and machine learning workflows in the age of LLMs.
[00:42:42] Producer: Watch now at snowflake.com/build.