 So the reason we are discussing going into the history of the direct benefit transfer program and its fruits in the Andhra Pradesh smart card project is essentially to understand how technology systems have evolved in India, primarily because if you look at the history of other or how technology is being implemented in this country in terms of standards of when you bring in biometrics, when you bring in nitrous, what are the actual origins of this, right? Like we are not essentially sure why these systems are popping up, but if you trace back into it, a lot of this is emerging actually from existing standards, international standards, existing pilot projects that were potentially implemented across the country. In particular, the usage of biometrics and banking actually emerges from the Andhra Pradesh bank card project, which kind of came in because back in 2006 when early UPA1, you had the NRGA system just rolling out, people just figuring out how to start giving work to livelihood to these people and start actually paying them subsidies. And you actually see a bunch of these things create problems for the government in terms of wage distribution, in terms of duplicate people, I mean sure there were issues in terms of multiple people who were trying to benefit. There could be political scans which were probably created by the local leaders, but essentially created a bunch of issues in terms of distribution of subsidies. And the issue in terms of distribution of subsidies was kind of always there, especially in the public distribution systems where people were trying to draw out extra rations. And states like Andhra Pradesh and Karnataka have been trying to deduplicate them or find duplicates who were misusing these systems for a while. So Andhra Pradesh and Karnataka I guess were one of the key early states which were doing this because both the states were trying to use e-governance or IT in governance and both of them were equally trying to do things. But it was mostly Andhra Pradesh which in this particular case has been documented. So what AP does in 2006 is they should pilot project order to essentially start transferring NRGA and social security pension wages electronically. At that point of time I believe the government employees were already receiving their salaries using electronic transfers because the electronic payments were already in place except it wasn't easy to do electronic payments for people in villages because they really didn't have bank accounts and forget accessing electronic bank accounts. So and this was a time when there was a lot of effort to create financial inclusion to get people into formal banking system for financial inclusion programs something which NABAR was essentially trying to do and there was this financial inclusion technology fund under which is what this entire project of pilot was tested. The pilot notification actually has a detailed list of things that they wanted to do because it's a pilot, it hasn't been implemented. They outlined the plan and the list of agencies and entities which would be coordinating with the pilot. The requirements definitely included identification of beneficiaries because that was one requirement in distributing subsidies. And to develop this they actually worked with the Institute for Development and Research Banking Technology which was the tech body for banking agencies which was trying to experiment with banking technology systems. So at some level the idea to start using electronic banking or a kind of a direct benefit transfer originates with the AP smart card projects. The reason biometrics were involved is essentially because of the standard of the smart card standard called SOCSTAR, a C-O-S-D-A smart card operating system for transport application. SOCSTAR as a standard kind of emerged because the Ministry of Road Transport and Highways wanted this ID card which they essentially wanted to use to start giving driver's licenses. And this was a point of time they were also having issues with people using creating fixed driver's licenses so they needed some kind of verifiability of what is an original license and what would be a fixed license. So the best thing was to start using a chip biometric based, chip based ID card. And SOCSTAR was developed by NIC, ID Kanpur and CDAC was also involved. And it kind of emerged around 2002-2003. This is the point of time when India was also debating to getting a national ID card. So the Multi-Purpose National ID Card Subcommittee at the Ministry of Home Affairs also recommended that they start using SOCSTAR as a standard for the national ID card scheme. So the idea of biometrics in banking or whether in ID cards kind of emerges from this standard because this was the standard ISO 786 or the ISO 1443, the contactless cards where the existing standards at that point of time to actually be used for ID cards in general. So this actually became the basis for every ID card, whether it's subsidy, whether it's driver's license, whether it's ID card in general, there are plans to get this for e-transport, so e-passport and even pan-pack. If you actually go look at the SOCSTAR.com website, you'll find a lot of documentation including source codes and technical documentation of how they were basically essentially storing just some values of any identity information that you wanted to show, whether it was for driver's license or MNIC or PDS. So SOCSTAR kind of became the base standard which promoted biometrics at some level. Coming back to the Andhra Pradesh project, so I'm actually using slides from 2007-08, even though the project notification kind that came out in 2006, it took them a while to coordinate with various agencies including the RBI and PANTS to essentially implement them in two districts of Varangal and Karimnatar of first world Andhra Pradesh, it's telling us now. So they decided to use these smart-pack systems in two schemes, the social pension schemes and the NRGS scheme, and basically to actually get this rolling, the RBI actually appointed a couple of committees which were looking into usage of information technology for financial inclusion systems, which the RBI was very closely monitoring because it's a pilot project on electronic payments, and what essentially it led to, I mean, if you ask what did the pilot show us, they had issues in terms of people not turning up. So there were around 474 people who never really turned up to claim NRGS payments or pensions, and there were also people who were enrolled but they did not turn up. So all of them were termed brokers. So at some level this is the origin of the savings that claims that they always use, that using biometrics or using electronic benefit transfer could bring down the cost of implementation and also save money. It kind of emerged from this pilot. At least this is the first non-documented case of it. So the RBI committees on financial inclusion are interesting because if you look at standardization of technology as a process, if it started with SOXTA for the identification cards, in terms of using biometrics for financial technology or using it inside payments, industry kind of emerged with RBI's Comptient Financial Inclusion, which heavily looked into this biometric data exchange formats, the ISO 1974, so 1995, because at that point of time this was the standard at one. Again, everyone was using, they were looking at global standards, so you find the two L and ISO standards and you're just trying to adopt them. This is where you try to standardize how the biometrics will be collected and what kind of biometrics will be collected. The idea of phase image data and ID data essentially also emerges from the RBI Committee on Financial Inclusion and most of this was standardized by this committee. Even though when we go back to SOXTA, the SOXTA standard is, again, an equivalent ISO standard, which an IC can just adopt it, but it was actually tailor made, like you had to write the code for it to share it with different ministries. In case of the RBI Committee on Financial Inclusion, they just issued a standard. It was in the RBI which was essentially deploying it. In fact, it was the private companies which were implementing this system, so you had companies like Fino, which were essentially collecting biometrics, issuing these smart cards and at the same time having the post machines where one could verify the person who's coming to collective wages or who's essentially there to collect the electronic transfer of money is not actually a duplicate person. This idea of using Fino or banking correspondence as they call it, the model emerged much before this pilot project because it was something that the RBI was trying to promote for financial inclusion projects in general. So when AP starts to look at this project, they actually adopted whatever was the best practices and they said we'll adopt the business correspondence model. The biometric standards in banking while they were framed by the individual committees, what essentially happened is it was this RBI Committee on electronic benefit transfer which actually studied the entire Andhra Pradesh smart card project to finally arrive on how to actually do electronic benefit transfers. I think this was around 2008, while you had the biometric standardization for financial inclusion projects by RBI in 2007, AP announces its project in 2006, but the entire pilot project to be implemented in those six months it took them a while. So it was kind of evaluated by the National Institute of Smart Governance which actually produced a report on it. This was kind of an issue in the first and second meetings of the committee which was looking into the electronic benefit transfer program. The issue in front of the committee were basically different. One was a business process issue on how the bank should be distributing this whether it should be a bank branch model, whether it's a bank and business correspondence model. So they were evaluating not just the technology systems but essentially the role of banks and role of other institutions in this entire process. And this committee on electronic benefit transfer is actually what recommended the smart card biometric identification system that Andhra Pradesh has implemented. And this is where the idea of deduplication kind of emerges because as I was showing you again Andhra Pradesh when it actually implemented its pilot it had issues of people not turning up or to claim essentially same people with potentially different ID cards essentially different when I say different ID cards you could be the same person but you're trying to enroll into service programs using say a bank using a ration card or using a voter ID so that various IDs that were eligible for you to enroll into any of these schemes. So they needed a single ID to be used everywhere without a single ID there was a potential duplication of systems where any individual could actually be enrolled twice and to evade that the idea of deduplication check kind of emerges if you are asking how were they planning to do deduplication check they actually realized this was a very computing intensive task and they didn't really want to have a centralized database right like if you look at the whole idea of the SOPSTA standard card it actually stores the biometrics on the chip on the card like the biometrics are not centralized so the committee actually says we should be deduplicating people at a restrict level but it's often in other case studies in other implementations it's been found that there could be people who was potentially claiming subsidy in two different districts and a district level deduplication check was kind of not really a viable solution but this is what the committee offered at that point of time due to the technology limitations that were available like you couldn't really do an entire national level deduplication check but the idea of deduplication primarily emerges from this committee and it's kind of interesting that this committee evaluated only the Andhra Pradesh project even though there are reports of similar projects that were being carried out in other states as well like you had Karnataka which is also trying to use biometrics trying to do deduplication because doing fashion delivery I guess because RBA was trying to limit itself to the concept of payments and not necessarily looking into the whole idea of subsidy delivery in all other schemes they kind of limited their scope to just the electronic benefit transfer where they're looking at how could we start distributing wages for NRGA and pension where it's actually money and it's not like a grain that's being given out in case of BDS the evaluation of the smart card project actually happened because National Institute of Smart Governance actually recommended that the system be evaluated to the committee so first meeting of the committee itself Andhra Pradesh government actually recommends the whole public-private partnership I mean you're technically using the public-private partnerships when you're using business correspondence pre that, like before you want to introduce electronic benefit transfer schemes most of the wage distributions was actually happening through India first so it was mostly being done by the government because banks were technically not available like physical bank branches were not available in remote parts of the country where NRGA work was actually happening so it was near impossible for a private banking company or even for a government banking institution to start distributing these wages at these remote villages so to distribute the NRGA base payments India first was the essential point of contact but Andhra Pradesh actually starts recommending using private sector because they had issues using Indian force right like it was a huge challenge to distribute money to potentially pros of people who were actually part of the scheme and all of this happened in one of the first meetings of the committees there in the committees report if you look on the RTS website the first meeting also essentially said NISG at that point of time when the committee was initially meeting wasn't really sure what this system was working so they actually recommended the technical and business evaluation of the system this is when you have the RBI and NISG together evaluating the system so there are multiple reports like the RBI one so the NISG report is also really interesting because at that point of time they actually okayed the entire technology system because biometrics and smart cards they were fairly adopting ISO standards there isn't anything new that they could be suggesting but NISG was primarily looking at the business practices and processes more rather than the technologies if you actually look at the report there are questions looking into how the data storage is going to happen what will be the security arrangements that will be possible I think I ended the screen sharing story I'm just going to the report of the NISG report so if you look at the NISG report it was fairly interesting that they were looking at all crypto related issues what kind of public infrastructure how will the crypto be put on mobile like you're using an NISG security chip which is 72K is our core processor smart card this was high back in 2006-2007 so these were the security measures that they were looking at but they also had questions on to mostly Fino which was looking at implementing the systems there are essential questions like what are the cost impacts of using contactless card versus contact cards the technical evaluation as I was saying is like they really looked into all of this what would be the security issues what would be the cost memory cost what would be essentially how yeah I mean they were even looking at the legality of these issues the IT Act does not mandate PKI to be used in smart card solutions so you were essentially evaluating every individual component of the system the idea of deduplication identify definite requirements of deduplication process along with the cost involved was also evaluated it wasn't easy to start doing deduplication processes for a lot of people like you're talking about millions of people in the country even at that point of time into 2006 but NISG actually okayed the entire system NISG actually okayed the entire system by just suggesting few basic changes to business processes coming back to this question why are we looking at these things what is the actual need of even evaluating these systems like when you think this is kind of the prehistory of other the other project kind of takes over from here and when we criticize other for its various failures it's not necessarily that the failures were first experienced during the other project even if you actually look back go back and look at some of the documents from the IT smart card project they actually had issues of people who were physically challenged to couldn't really access their subsidies so the idea of introduce a concept like somebody else will share their biometrics and collect your subsidy your wages was essentially already introduced in Andhra Pradesh way before this was again implemented in other and a lot of this kind of evaluation that we have seen that has happened with the AP smart card project kind of really didn't happen with other or even if it did happen they were reports by UADIA itself which were trying to recommend things like biometrics, idols but most of those origins kind of go back to these committees and reports and I frankly don't know whether at this during 2008 2006 to 2008 whether there were any public engagement on these issues like when you actually started looking at UADIA definitely they were not so open at the early stages but eventually they start holding public consultations so the evaluation of these standards and creation of standards and technology systems kind of makes up an important part for any technology system to scale up and this has been happening and the AP smart card project is essentially the root of this but I think one thing if I have to say now maybe everyone has probably overlooked at it in 2008 is what happened with errors like people who physically challenged people you really can't evaluate these things in a small pilot and this is the kind of thing that they also encountered with implementing the other program and they eventually started like a better fitting things but this idea of identification and biometrics not necessarily originates from other it's always been part of our financial inclusion programs and other public distribution systems because they were always broken from the start open it up with questions the idea in trying to bring up this history is essentially to even look at how we start evaluating current upcoming standards of systems which are ongoing just allowing everyone to talk if anyone wants to ask any questions you can just unmute yourself if anyone has a question can I explain that just historically I guess so while this was one side of things we are going to look at actually look at raw data of the talent benefit transfer scheme like next this Saturday part of this cashless consumer meetup so I've been filing a lot of RTIs trying to actually get data on transactions and transaction failures both with other and without other across the country states and centre for all the schemes with the DBT mission and they kind of supplied data for the past 2 years from 2017 to 3 years have data until June 2020 it's actually on Karna's Github account if you just go to github.com slash Karna you can find the DBT data which is already up there we will be looking at the technical design of DBT and why failure kind of happens in DBT delivery currently and what are the actual numbers I know all of us kind of debated this at length during the other judgment or the other challenge in Supreme Court but there has been no actual evidence based analysis of these systems even if you look at the errors that happen at UIDKI they are completely different from the errors that would happen at NPCI because of the number of institutions that are involved in this process okay so Ambika has a question could you tell us a bit more about how they dealt with the errors in AP pilot you mentioned this I think might have missed this was deduplication of success in that context okay so I don't think there was deduplication in AP deduplication was recommended by the RPA committee after looking at the AP project but what happened at the AP project was identification of beneficiaries with some level of consistency I won't call it that but at least that's what they called it right like so you're trying to identify people accurately uniquely this whole idea of uniquely identifying people actually emerges with technology emerges from here it's not like people haven't tried identifying beneficiaries with relation cards or elsewhere I mean if you really want to understand the historic context of ID documents that recommend you actually read Tarangini Sheraman's book Identity Documents in India but in terms of usage of information technology in ID cards kind of goes back to SOCSTAR and it kind of comes back to the AP smart card project when you're looking at dbt I mean if you actually want to look at issues in terms of ID related issues I think the SOCSTAR website has a fair amount of details but so the question on terms of errors right like there were a few issues I think this was looked into heavily in the NSD evaluation still smart cards create any errors at all right like technically it's a card you're using the POS device to verify the biometrics now the issue could be the biometric fingerprints may not match we have no data on this at that point of time but there are documentary evidence of them facing issues I can share more documents I mean if you actually want to look at some of the documents on the AP smart card project a lot of them are on archive.org on the way back machine actually you just have to go and type in haven't shared any links here but all you have to do is Google development department of AP it's actually rd.ap.gov.in on the way back machine and go back in something like 2008 or 2009 you will find a lot of these reports even the NSC evaluation report is kind of not available anywhere I mean most of these things are not in current public websites I'll be mostly uploading them and I can email you trying to gather as much information as possible as well but there was definitely that question during the pilot whether errors can happen with biometrics and they felt it was minimal definitely they didn't look at the kind of research that went into other post patient of UIDI definitely did not happen with the pilot project it's also in some ways you could say that some of these reports and standards that are made don't really look at the complete profile of these systems but they are just recommending these systems based on small pilots and it's often the case in all technology systems anyone else is interested to know a bit more about the technicalities of these systems I think a fair bit of of things that went wrong with others already went wrong with this project in these official reports because when the RBA was looking at the evaluation of this project I don't think they carried out their own independent evaluation they actually asked and I have seen the work of that so you have you always have consultants and these organizations which may not be really putting in all the issues that I face and when you are talking about bringing technology into public systems often the public part is your consultations with public feedback from the public kind of never really happens at the way it should happen yeah I think that's pretty much it has more questions I'm happy to also take some feedback what you want to hear about the current GBTs the upcoming session on Saturday Srinivas hi this is Harsha here yeah so working in Telangana for example we have the famous Raitha Bandhu scheme where you know probably the government says it can actually credit accounts for like 2-3 days to all farmers who are eligible and understanding how the government works and also trying to discuss with a few collectors and administrators this time there is a lot of push for authorization totally right wherever there are gaps a few thousands in each district they are also pushed in to make sure all were authorized their accounts and the amounts are pushed in are in itself and then we have the whole cash benefit the DBTs which are done so what are we actually finding what is the way you know as a civil society forward is also I was trying to think through and are we looking at any other better models in this aspect different countries also in the similar aspect for example just to also I think Sunil from Sunil Abraham was at one point of time saying instead of other we can have a card which is you know so are we looking what are the alternatives we are looking at when we say about DBTs and things is also something else you know trying to yeah yeah so I mean the idea of using cards instead of biometrics is always push right like Sunil at yes the internet and society was always talking about using Soxta in fact when he called Tom Smartcard he was actually the footprint of the Soxta Smartcard standards because at that point of time they were already standards why is the UIDI collecting biometrics at that scale for the whole database and in terms of security issues in terms of other challenges they were always recommending that let's start issuing cards instead of biometrics but a lot of this has nothing to do with the electronic payment transfers a lot of this is about just the identification layer in fact when you look at wage payments, salary payments that happened in Andhra Pradesh itself they were not about IDs of the government employees they are mostly based on the bank account numbers so all you have to do is just transfer to use something like NEFT to start transferring money directly into accounts in fact usage of NEFT for NRGS was always done even now there is a significant portion of dbt money which actually gets transferred from other based electronic transfer methods like NEFT so other kind of became mandatory because it was pushed for this whole dbt exercise there are frauds and all of it people are skimming money and all of it but I think at the end of the day the ulterior motive of the private sector was trying to basically sell to gain money the 2% tax interest that they get for every transaction but frankly I think at this stage I would even say smart card is also a waste if you are talking about actually doing electronic bank transfers you just need a bank account you don't need the ID cards other kind of complicated all of this because of the whole biometric authentication verification and failures at that stage I mean if you go look at UIDA's website UIDA itself says they have 99 types of errors in which the whole other authentication can fail from internet marketing to biometrics not being a bit of a nice there are like around 100 error codes now when you start linking that to POS device which is your android or any tablet with biometric sensor it kind of increases when you actually go into start using banking on top of other number in fact it goes further we can't compute them because UIDA never realizes this data but I am going to be looking at me and Shrikan Plashman will be looking at some of this with actual figures on data and how much errors are happening did they go down in the past 3 years did they go up both at the center and state level in fact the data is already up online as I was saying you can just go search for github.com and you will find the dbt data it's just excel files have a look at it the issue still exists the data itself shows us that they are trying to work on it but again data always doesn't capture everyone's issues and if you are talking about alternatives at this stage I don't think you can even listen to it but thankfully because of COVID everyone is shutting down usage of biometrics for delivery of subsidies especially for pds so I think they are just giving people are just giving distributing some of these subsidies by identifying them in other ways