[Music] hello everyone welcome to another episode of tech unhinged we’ve got a very special guest here today with us uh Dr Amar Malik honestly introducing someone as accomplished as Dr Amar is a bit humbling while he’s out there changing the world I’m just trying to keep my tutu list in check so I mean he’s written more books and research papers than I’ve probably read in my entire life so rather than fumbling through is Incredible Journey I’ll let the expert do to talking Dr Mar welcome on this show thank you very much zaro uh please share a bit about yourselves your work and how you ended up at uh Bill Gates and Melinda foundation and we’d love to hear the whole story thank you very much Zara it’s very nice to meet with you so Zara I began my journey um in uh in laor at lams about 20 years ago I studied an economics degree and I ended up doing a PhD in public policy from George Mason uh much of the focus of my research at that time was on cities and urbanization and I was fascinated by how big interventions from governments like the laor uh Metro Bus project which was started in 2014 how will that affect different groups in society so how would men and women and children and families and Working Class People how will they benefit from that and what my research showed me at that time was that the benefits of the these interventions are not equally distributed they are actually quite unequal and then I went on a quest to try to explore why is it that when governments bring infrastructure to populations that the underlying uh inequities between the Haves and the Have Nots how does that play out in terms of the benefits of those so that really became a long-term quest of my career and ultimately I ended up working for the last four years at a research lab called a data which is a um at the College of William and Mary and we used to produce the biggest uh most comprehensive and detailed data set on China’s overseas development Finance because as you know the Chinese have in the last 10 years in particular uh built quite a bit of infrastructure all around the world and the same questions that I was studying for the case of laor for my PhD work became relevant in not just our own country but in many other countries around the world and today my position at the Gates Foundation is related to uh the uh development financing uh architecture which means uh rich countries giving money to poor countries how much is Flowing um what is it for how effective it is how do we measure it and how do we know what is working and what’s not working of course this is happening in the context of uh a global crisis on foreign uh foreign debt or sovereign debt which means a lot of developing countries are now having trouble repaying the debts in our own country in Pakistan I think something like 60% of our expenditures from the government side are just in paying the interests that we owe to to um to different banks around the world uh which means that the government really has very little capacity to spend money on things that really matter to the people of Pakistan such as schools and uh hospitals and so on so that has become a long life uh quest for me and um I’m very happy that I have this platform uh to share some of my thoughts with you uh that’s absolutely uh fantastic Amar um you know uh uh from your um introduction I can I can ently say that your expertise they span over some fascinating areas but since today uh topic we are going to focus on foreign aid and how foreign aid can be paired with modern technology and what is the impact of uh uh foreign aid programs being paired with uh technology so can you just uh give us a bit of insight into how you see uh technology transforming the way foreign aid programs are being run delivered and my year today so maybe before I answer your question specifically I can spend a couple of minutes explaining what foreign aid is and why it matters so as you know in I’ll take you back in the 1940s when um there was second world war which created a lot of Devastation particularly in Europe but in other theaters as well in Asia there was a lot of fighting that took place of course Japan got destroyed so after that uh the United States led the biggest reconstruction ever of Europe called the Marshall Plan and after that also the United States since it came out as the Victor together with its European allies set up a whole system of Institutions which we call the Britain Woods institutions which includes the international monetary fund which stabilizes economies and of course everybody in Pakistan knows what the IMF is the World Bank and later the Asian development bank and others the idea was that rich countries because they have enormous resources and wealth uh on the one hand have Tech technical expertise and the knowledge to help uh developing countries um in Africa South Asia Latin America there was lots of developing countries in fact the majority of the world’s population lives still in developing countries that uh they could help just like the the Americans helped Europe rebuilt uh after the second world war the same could be applied through these institutions to the rest of the world um in other words rich countries could share their expertise their money and their knowledge with poor countries to help them improve their governance system and for them to develop and give people ultimately the objective was to give people a higher uh standard of life so that um created what we know as the foreign aid industry or the foreign aid system where in the 1950s and 60s uh that time the richest countries of the world mostly European countries but also Japan and Canada and the US they made a commitment that they would commit 0.7% of their GDP or global national income into uh foreign aid now that that was a commitment that was made to them uh based on economics ideas that uh certain amount of capital money has to be transferred into developing countries in order for them to develop now what foreign aid has taken is many different forms uh it could be anywhere from uh a school education training program uh to uh public diplomacy sometimes uh rich countries uh bring uh scholarships programs like the full full bright scholarship program that we run in the US runs in Pakistan that’s an example of foreign a uh foreign a can also create uh hydrop Power projects like the mangla dam and the Tela Dam today China is building a huge number of Highways and bridges and tunnels in Pakistan so uh the the idea is pretty much that you as a developed country would provide assistance to a developing country to do all sorts of projects to help them improve their living conditions now having said that like everything else in the world foreign aid or this industry has also benefited from the Advent of new technologies of course um the the one aspect that I know a lot about is foreign aid data and the evaluation of projects so when I say foreign aid data it is not an easy exercise for us to keep an eye on the total flow of development financing from Rich to poor countries across different sectors in different geographies so the systems that we put in place to extract that information from the web from Twitter from official Source documentation from statements from media reports and articles and putting them in the same format across the different owner countes so we can compare who’s doing what where uh becomes incredibly important so I would say that technology has really enabled us to get better at tracking uh the flow of financing and that really makes a big difference on the ground for example if I am the government of Germany and I I want to invest in the Health Care system of uh Zanzibar which is a uh province in Tanzania it would be hugely beneficial for me to know which other foreign donors have uh already provided assistance to the government of Zanzibar and if they have provided assistance what aspect of Health Care Healthcare have they provided assistance on is it malaria is it hospitals is it x-ray machines what is it so that the interventions that I bring about then become complimentary to that I don’t want to do the same thing I don’t want to go to the same Villages I want to spread out the benefits and make sure all areas in that Province are benefiting from on that now part of that is just knowing how much money is Flowing uh who’s receiving the money on what terms was the money given out was it a grant which means a gift or was it a loan uh do you have to repay that if it was a loan how much of it has been repaid where were the projects implemented and I think one of the most exciting areas where we have seen massive improvements in foreign aid tracking has been in geospatial uh aspects which means where are these projects being implemented so knowing the where and the when is incredibly important important to be able to measure the impact so for example if I know that Village a in Zanzibar received a malaria clinic in the year 2016 uh right before the rainy season started I can easily track the disease cases from hospitals uh where malaria cases are reported every year to compare the before and after effect of whether that malaria intervention worked or not if it worked how much did it work did it work as much as it did in some other country or did it work less so all of these critical questions can be answered by the use of technology so that we can evaluate the effectiveness of foreign aid programs only after evaluating their effectiveness in this data and evidence-based manner can we make a judgment as to whether um this intervention worked or it did not work and of course there is only a limited amount of foreign aid dollars available in the world so it is very very important for us to know whether the dollars that we are spending are actually benefiting um the populations or not and that’s where technology comes in so basically from what you’ve said I’ve understood that technology has basically increased transparency within foreign a and uh how do you see with the Advent of AI and ml how modern technology is going to reshape uh data analytics and the process of collecting uh information about foreign aid data so already you have some streamlined processes already set up for that so with the Advent of AI and ml is that going to change significantly or is there going to be uh uh lesser impact on that I think I’ll start with machine learning because that’s where the use case is much better established um already I mean AI I mean we’re trying a lot of different things and the benefits are of course beginning to show up but it’s still a lot of work to be done in the machine learning domain I think I’ll give you a simple example you know when we are uh a data used to collect data on what China’s development financing looks like in a developing country um let’s say uh China made a loan to Uganda to uh for $1 billion to to upgrade the Uganda entay airport which is the international airport in Uganda now that was a loan that was given out by China exim bank and the loan had some terms and conditions so it was a 15-year loan for example or 20-year loan um let’s say it was a 3 and a half% interest rate there was a grace period of three years after which payments would be due um and then there would be management fees and these fees and that fees so we have you know these data points let’s say we got our hands on the original loan contract like we did at a data and now that loan contract spells out all of these terms and conditions now what we would do is we would basically extract those key pieces of information around the interest rate and the term of the loan Etc inserted to our system and that system would then meaning like a like a information system like a form on a website where we would insert those pieces of information and that it would automatically populate uh many other fields that otherwise we would have done manually or through Excel sheets um that is one very easy case in which like hardcoding some formulas can help us complete um the collection of data uh and then what we did also was we put some uh data quality assurance checks so for example if uh the interest rate was uh 3% and we know that for this sort of projects uh other countries and China have given loans which are always above 6% and this is somehow at 3% that would be a flag that would be tagged to for a human to go and check why is it that this interest rate is so low could there be an issue with the data so the quality of the data gets improved quite a bit and we can then compute quite a bit of uh Downstream you know variables and effects of that another application of giving the same example was you know what happen to uh the efficiency of the Uganda ENT airport after the project was completed so we have now data on when the project was announced the date when was it started when was it completed so after it was completed uh we would then be able to see in the vicinity of the airport did you see any greater economic activity could you detect from satellites that more uh houses were being constructed uh maybe because there’s now more flights coming maybe there are more jobs created maybe there’s a mall that was built around so by using satellite imagery and and extracting high frequency data out of uh that area we could detect whether that intervention actually had an effect or or it didn’t have an effect so that’s I think a very easy to understand example of what I would say simple machine learning or simple checks um and for more advanced applications uh what I’ve begun to see is very good research coming out now in the area of uh sentiment so uh there is a research program at the University of Michigan called G delt where they have this enormous gigantic data set which updates every 15 minutes where they extract text uh they do text mining on hundreds of thousands of media articles podcast transcripts YouTube transcripts and stuff and then they perform some uh sentiment analysis on them or you can ask their machines to perform sentiment analysis on it I did a paper recently where I was interesting in interested in understanding whether uh after China or the United States starts to spend more money in a developing country for development whether the media of that country uh begins to give more positive reports about that country or not so the sentiment about China in Bangladesh does it go up or down after China gets more heavily involved in building projects I did the same thing for Pakistan and what is really interesting is that what you can do is you can sort through uh millions of records and you know when those records were were created so all the Geo News and Dawn articles that have ever been published over the last 20 years sort them out and it would tell you whether the overall sentiment of an article which was let’s say about China Pakistan economic Corridor you can sort by CPC and if it was a CPAC article has the media sentiment gone up or down and what you can do then is you can then correlate that with the implementation of the different projects and if you are more sophisticated you can even try to distinguish between whether this article is related to The syn Province or Punjab Province and you can see if the introduction of new projects by China or another Power results necessarily in building Goodwill for that country in um in the developing world so why that matters it matters because we all know that as much as we like to think that rich countries are benevolent that they want to do well for you and your country they are also doing it for their own benefits right they want to build Goodwill they want to make sure that their country is seen as a positive force in the world and that they are winning over the hearts and minds of people who are who are living in recipient countries right so for that aspect the media sentiment is a proxy measure or a marker for whether uh a country’s population or its media is looking at a rich country more favorably or or less favorably and then of course you can do same sort of sentiment analysis on Twitter feeds um and on other kinds of um things that are coming out uh in terms of like Vlogs and Etc so so that’s just one example where I think AI is beginning to make a difference but I’m sure there’s other applications that I don’t know of yet or I haven’t thought of so that’s pretty interesting so does sentiment analysis really have an impact on who gets the foreign aid of if if foreign a gets approved for a particular country or not well it is I think it depends on it’s a chicken and egg problem so does the introduction of foreign aid in a country create Goodwill for the donor or does the donor only go in those countries where it already has Goodwill I think that’s the the question and so what we have found is that in the in the belt and Road case China has literally gone to all developing countries in the world so it doesn’t really matter whether uh you know it already had Goodwill in a country or not however what we have found is that when China does get a begins to get a positive sentiment coming um it double down doubles down on a country so it gives more uh attention to the media its embassies try to engage more um with the media they would then fly journalists for training programs to China China and by the way this is all things that British and American uh and other European powers have been doing for the last 50 60 years so China is not coming up with something new it’s just that it tells you because I’ve looked at China very carefully it tells you that when things start to get better and better public sentiment and media sentiment improves then China gets more encouraged and tries to do more and more and more for that country and the opposite is also true that when things start to go bad when the impression is left that China’s given us so many loans that we cannot keep up with the payments Etc then China tends to lay low and let things move on its own so it’s a it’s a double-edged sword but I definitely see a correlation between media sentiment public sentiment and foreign aid so now we’ve already talked about the impact of AI and ml into uh you know in enhancing foreign aid in different ways so do you also see some gaps there where you think you know AI has not yet been leveraged and there are gaps and inefficiencies still present within the system uh where we can effectively uh leverage Ai and other advanced technology to sort of uh uh overcome that that’s a great point I think that there is still a lot of Under reporting of uh development flow or financial flows of all sorts and when I say Financial flows I don’t mean just foreign aid I also mean foreign direct investment I also mean remittances which is the flow of money from expats who are working overseas into their developing countries I also mean tax collection in developing countries so if you think of a country like Pakistan or Bangladesh for example uh the amount of money that is spent on public health is not just spent from the government’s own resources although that should be the main source which is coming from taxes it is also spent by foreign aid Private health um I think there is still a lot of blind spots in what we know about how much money is actually spent by the people of Pakistan and the government of Pakistan and foreign agencies on the health of pakistanis so anything that we can do to help improve the collection of data on let’s say the healthcare sector whether it is by looking at uh Bank transactions or linking Bank transactions to people’s behaviors in terms of spending patterns uh establishing better data collection systems in hospitals and making sure that they are more digitized so that we can know why the patients are being taken on for example this week we are beginning to see media reports saying every day how many thousands of pakistanis are going to hospitals because of smog they are going for respiratory problems right so it’s a very important data point if you don’t shine the light on a problem uh by not looking at the data you will not see what’s happening so we’ve had smog every year for the last many years for the last 20 years but if we had good data on how many people are being admitted to our hospitals uh during this season for respiratory problems and that we can compare it to an average Baseline um that would be very powerful so then you can go to Rich donors private foundations or The Who and say look how much economic and human cost this smog is having on on the people of Pakistan unless you collect that data unless you know what’s happening and you connect it I think it’s it’s it’s impossible so that to me is the biggest blind spot and I think that machines can help us uh do a better job in collecting this data standardizing it uh and then analyzing it for uh the welfare of of the public yes I absolutely agree with that viewpoint but the problem is that with government organizations and government institutions they generally uh you know are not that receptive to adopting new technology so do you see that sort of changing in the near future or CH uh seeing a shift in the trend look so when I say that I you know there should be better data collection more data collection we should know more about this I absolutely understand why governments would be resistant to change I know in many countries uh there are laws about privacy for example I live in the United States where um believe it or not they still use fax uh technology to tax letters because according to American laws or some regulations or health laws they are not allowed to email me my health information so anything that contains my health information or children’s health information in particular they just don’t uh transmit it in an in an inse unsecure way and somehow fax is more secure I mean who even uses fax these days but this is an example of where laws have perhaps not keeping keeping uh kept up uh speed with new technologies somebody was telling me just yesterday that if you hire somebody to work for you in Germany the German law currently doesn’t allow us to use uh digital signature Technologies like DocuSign so you have to physically get like a a contract mailed in fed from FedEx over from Germany to the United States to physically sign it and then to send it back for counter signature uh America in America that’s not a problem we can do it all we do it all the time so my point is that the local laws and regulations around privacy U need to be updated and they have to keep Pace with the developments of the Technologies of course in developing countries the situation is much more complex where um the governments are so embroiled in some of the most basic things like protecting people’s security or making sure people have enough to eat and there’s hospitals and children are going to school that the issues that you you and I are talking about might feel like they are second order priority issues that this is not such an important thing but my point is that uh I think that uh by keeping uh the regulations around data privacy uh and data collection and data reporting by keeping them up to date with where the reality of the Technologies is ultimately it is only going to benefit the people uh of of those countries and I think it is in the interest of the people and the government actually uh who is spending all this money um to do a better job in collecting this data if they don’t Target their foreign aid or local Aid or foreign aid programs better then they will waste a lot of the money and and in fact we have some good examples in our own country with the benir Income Support Program during covid there was good targeting of smart lockdowns Etc where data is being used to improve the efficiency of government functions uh I think that there is resistance to change with everything and with technology since it’s so unknown there is also generational problem people don’t like people from a certain generation don’t understand these Technologies or the benefit from it I think all of that plays a role in that but above everything else I think it’s the law that needs to keep up with it and that means that we have to work with parament Arians we have to make sure that there is awareness about the benefits of uh Technologies and I think what we’re doing today by way of a conversation about the benefits of uh this kind of data collection for the Public’s welfare I think is a good starting point yeah thank you Mar that was really insightful so uh how do you suggest governments uh should look or hunt for a vendor or a technology partner uh what are the key factors a government should consider in order to find the right fit technology partner for them that’s a good point I think that uh I see that a lot of times governments are uh in developing countries are not very sophisticated in the way that they write out the requirements that they have for projects um which means that if they’re going to have um some data privacy protection requirements in new technologies they ought to spell that out fully and transparently at the start uh for example I remember that after Pakistan introduced the Smart City technology or sorry safe City technology with CCTV cameras in Lor it was only three or four months after the cameras had been turned on that the government uh realized that they they did not have proper uh protocols around handling of that data so if the data from a camera captured a crime under what law uh or what procedure would uh Punjab police share that with a court of law um and they had to stop the cameras for a few months and then went back to the Punjab assembly got a new law passed they did not think through that and they had to bring foreign uh uh vendors to try to help them build the law and the technology from Europe so it’s an example of where the government did not really think through uh that and of course in selecting the vendors they also perhaps did not take into consideration that vendor’s experience or ability to maintain data standard that they they need um I also think that based on what we just discussed today about need to mine this data for an analysis and to do a better job in targeting foreign a or any projects for that matter that dat have a garbage uh in it you’re going to get garbage out but if you have data that is being collected and transformed in ways that are standardized across different projects uh I think that could have huge benefits so I think those would be key considerations for the government uh but of course I know that in reality cost ends up being a huge part of what they’re they’re worrying about so uh before we wrap this up uh what would be your uh suggestion to technology vendors that if they are to secure uh government projects and they if they are to work with organizations uh that help foreign aid and work on data regarding foreign aid and the financial records regarding that what are uh the key things that they should have or Implement within their organization to secure such projects in the future especially for companies that are located in Asia overall especially Pakistan and India I think you should invest uh a good amount of time and effort in understanding the the data privacy regulations um especially this is true for Europe where uh EU gdpr is a huge thing and if you don’t know those requirements and you cannot speak that language in uh even the proposal or with your client meetings I think it will be very challenging to get good contracts or any contracts because I think the perception in a lot of developing countries is that countries companies which are based in developing countries um or in countries which don’t have such strict laws about privacy that they not able to or equipped to do it um to to meet their standards um and this might be a myth I don’t know but at least taking the time to understand uh the government’s compulsions and stepping into the shoes of the person who’s giving you the contract and understanding what what kinds of things must they be going through according to their legal uh regulatory requirement would be good and I think the good thing is that at least in Europe and North America Government Contracting is fully transparent by law the governments are required to share the details of their request for proposals proposals sometimes even selection criteria and the evaluation methods that they use to pick their vendors is also in in the public domain so taking some time to figure out the correct uh data portals where uh your client agencies have previously shared data on who they picked why they picked I think doing that due diligence and then training your staff to be able to speak intelligently about those requirements uh I think would be great and of course the other thing you could always do is you could also partner with uh vendors who have been successful in uh doing this and then that could in the long term also help build your company’s capacity to be able to get contracts directly from the government on your own without them so I would say consider that aspect uh as well and then if and when possible I am a huge fan of open uh open source um code sharing and open source data sharing so whenever it’s possible in whatever format it is possible to share data and code and source code and methods and documentation of the code I think that would be very very helpful because you know one vendor can take a project that they’ve done put it online and another vendor can build something on it it would be efficient because you can take chunks of code and reuse it anything that we can do to promote this use of uh open uh open code open source coding open source data uh I think it helps create the public good that uh our people deserve a [Music]