 or greetings appropriate for the time of the day when you'll be watching the recording or watching the live stream from your time zone. We are back with another set of discussions around elections technology and the ramifications of introductions of technology into electoral systems. And so welcome to this session. This is part of our project. We continue to examine new technology introductions in election systems. You would note that this project is now called one vote. This is the same one that we had been hosting and promoting as part of the election at the Karana collective. And we continue to focus on how technology could work for citizens when it comes to voting and elections and all things related to that. So the way we take this forward is that we designed and post public deliberations such as this one to stimulate discussions around the evaluation and assessment of such interventions or policy decisions. And we are focused on creating easily accessed tools, frameworks and content which creates significant interaction and engagement. One vote project is hosted at hasbig.com. Hasbig.com as we would know is a platform for collaborations across practices surrounding technology, design, law, policy systems, data and related topics. These collaborations take place through user generated content that are shared by practitioners. The sessions at one vote are designed to enable participants to acquire the foundational knowledge and perspectives required to evaluate the intended and unintended consequences. of policy decisions related to elections and related to technology interventions. We bring a diverse set of perspectives which focus on identity, equity, privacy, security rights, agency and the socio-economic impact of particular persons. Now, today's session is a slight change from what we have usually done in the past, where it has been what we call a masterclass. Today's session is a discussion that is between those who are deeply following and keenly following the topic of e-voting changes. And my notes had that today we have in the panel, both as the moderator as well as the participants, individuals who are very, very involved and invested in this topic. But then I realized that not only are these individuals invested and involved, those who are participating today and those who will spend some time in the future watching this. We all are. It's just so happens that the moderator, Mr. Sunil Bhatri has been involved in various wide policy making decisions and he's had a good as a form of public servant. He's had a great insight into how things work. Srinivas, Kotali and Srikanth have been through various ways involved in ensuring that the policy around technology interventions do not cross the lines and limits of rights, privacy, agency and are continuously striving to various forums to be able to ensure that we do the right thing for them. Now the format of this conversation is very simple. I'll hand it over to Mr. Rajbir and I'll make myself scarce. We will use the raise hands option in Zoom to kill questions unless you really, really need to unmute and speak, which you can do and else you can type your questions in the chat. For those who are following us on the YouTube live stream, we will copy questions from the YouTube to chat into this channel and hopefully we'll try and get a conversation going based on how things roll out from there. So with that, I will end my spear and hand it over to Mr. Bhatri and welcome to take you for it. Thank you and all yours. Thank you, Sankarshan and hello everybody. Sankarshan in his opening said that there are intended and unintended consequences of technology or other policy interventions and that's very true. So I will refine myself to stepping back and asking some very simple things that we could keep in mind and then listen to the experts and learn from what they have observed and what they have to tell us. Now election, we all know, is the process of determining the choice of electors. There are three keywords here. Process, choice and electors. Let's look at them one by one. The most important I think is choice. The vote must be free, a free choice or cost less, but uninfluenced by coercion or inducements. The ballot is a secret for a reason. The voter is denied the option of even friendly assistance inside the polling booth. Why is that so? Because we can't be sure that the request hasn't been made under influence, the request to provide assistance. All that providing assistance isn't a cover for keeping watch over the actions of the voting. The polling booth is the sanctum of democracy. Inside that booth, for the time that it takes to cast the vote, the state guarantees freedom and independence of choice. That's not obtainable anywhere else, not even at home and certainly not on personal devices. The focus of our design should be to afford that level of protection and freedom from surveillance that's identical to that you find inside the polling booth. The next important word is electors. We want the electors, all the electors to cast their votes. If technology enables some who may not otherwise be able to vote, that's a net positive. But technology shouldn't make it more difficult for any others. Every vote is important. We know that. But randomly casting out a small number of votes doesn't affect the elections outcome. Technology's impact, however, is a little different. It's not always random. It may disproportionately disfavor some sections of population. And this possibility must be in our mind while evaluating the solutions. Further, technology may have other effects, some subtle, some not so subtle. For instance, who gets to vote for which candidate? If some or a large number of people from one area live and work in another, should they vote in their hometowns or where they work? Of course, this is a political question. But technology also could affect voting patterns and thus impinge upon what should be a political decision. Such effects need to be recognized. Our discussions this evening may throw up many potential concerns. And my hope is that the audience would be actively looking in the darker corners to find them and to throw some light. Finally, election is a process. We briefly touched upon the ways in which this process is currently designed for a neutral outcome. What we hope is a neutral outcome. It isn't perfect. But overall, the process seems to have worked so far. We may improve the process. But we should only do so after exploring the ways in which an improvement proposal might affect the outcome. We must try to estimate the direction and magnitude of its impact. Constructions of cost and convenience enter into the discussion only after we know these tradeoffs. I'm saying so because technology is often held up as cost saving or as convenient. But that comes later. Today's discussion is a process too. Its objective is to validate a design. But to validate a design that touches on the fundamental processes of democracy. And therefore, for democratic legitimacy, its validation must be useful and understandable to the average voter. Expert authority or any authority, accountability. The experts may have an agenda, which is fine, of course. Therefore, it is important that their submissions. We offered in a language comprehensible to the voters, while being open to challenge from other experts who may have their own agenda. If the experts don't communicate well, that is, they don't wish to acknowledge the concerns or opinions of the average person. Because of their supposedly superior knowledge. They undermine their own efforts. For an illustration, we could look at the anti-vaccine narrative. That enrages the experts and defeats their best efforts. So in all humility and openness, let's share our concerns. Let's learn from the observations made during the voting experiment in Telangana. And evaluate our opinions. First, and also those of the others. We have two expert speakers this evening. And everyone shall have an opportunity to ask them questions. After both the presentations are over. Please keep a note of your thoughts. Let us still write them down so that you don't forget the questions. And I'll look forward to the discussion. And invite Mr. Srinivas Kodali to tell us what exactly happened at Khammam on October 20. During the voting experiment conducted by Telangana State Election Commission. Over to you, Mr. Kodali. Thank you. Thank you so much. So I'm going to play a small video that I recorded. So what happened in Khammam was a mock election. So which is, I think is really interesting because mock elections are held for three different reasons. One is to, as an education purpose, like they do it in schools and colleges when young voters just turn 18. You're introducing them how election works. So they hold mock elections. Then the other one is also to introduce new formats of elections. So this is the second time where you think, okay, this is a new type of election. So we're going to hold a mock election. There's also a third one. The third one is actually to mock people in power. Like, oh, we can hold actually fair elections better than you. You really don't know how to hold an election. So three main kinds of mock elections. So we had the second time it was supposed to be to introduce people to how this new voting function. So I'm going to just share my screen, show a tiny video of the screen recording of the voting process. I hope you can see my screen. So it's a screen recording from a phone. So this is the app. This is the voting process. Once the app opens up, once you after you've registered, you enter your epic number, which is your vote ID number. And then you enter your phone number. So it kind of gets validated across their databases, whether this phone was registered with this epic number and this phone number itself. So when you do registration, they're tying your details to a particular phone. And then it gets verified again on the day of the election. So they're authenticating the device. So it's using OTPs and other device parameters to do it. After that, there is one more secondary check of identifying your face, but that you're a live person and not somebody else is voting on your behalf, right? Like you are the actual voter who is voting. So you're required to verify yourself using a video. So you have to ensure that your face is being in the exact box that they have. Once they do the facial recognition, once you're confirmed that you're actually the person who is voting with the epic ID, then you get this screen where you get to decide whom to vote for. So you've got, I'm going to pause it here. You've got alpha, beta, gamma, since it's a mock election, you don't really have actual individual names. So there's a ballot ID, the ULP where the elections were held, and there are four candidates. So it's a dummy elections 2021. So they're having a disclaimer because mock election. You can have options of different languages, whether you want to change the app into Telugu or Hindi so the information gets changed. I'm going to just play it again so you can, yeah, so this person has voted for alpha. After you vote for 20 seconds, you got a message saying there was a vote, so I'm just going to go back. So dear so-and-so, your vote has been successfully recorded in the voting, thank you. So you got a message and for 20 seconds there was a screen which was showing whom you have voted for. So the app which was designed doesn't really allow you to take screenshots by default. If you take a screenshot, you'll get just a black screen. I was able to record this video by connecting it to the phone to my laptop and doing a screen recording after networking the developer options. So there's some kind of security measures to ensure that people can't just take screenshots of whom they are voting for. Okay, so I'm going to end this back and come to again. So what you saw is the remote voting procedure of how people can vote for. Now in all ideal scenarios, this is like really great. People can just vote. It's a matter of seconds. You don't have to stand in queues. There's no other issue. Wherever you are, you can vote even during work. All you need is a 10-minute lunch break or something and you can do it. So I think this is what to turn out. So in theory, there are like so many of these advantages of having a smartphone-based voting. So the first time people tried doing this at a scale, not smartphone-based but electronic voting was in Estonia. In Estonia, you had a sliding smart card where you could use the smart card to identify yourself and vote online. And this was kind of an experiment that was carried in Estonia and several European researchers were really concerned about what Estonia was doing because Eastern Europe was kind of a laboratory where a lot of these experiments kept happening since the 90s. And depending upon how these experiments happen, they would be then exported to the rest of Europe or elsewhere. So the Estonian e-voting is kind of a standard when it comes to e-voting around the world. I don't think there are any other nation states which have done this. Now there is an active discussion to get e-voting in India. Within India, the election commission of India was working with the Tamil Nadu e-governance authorities and IIT Madras to actually bring a blockchain-based e-voting mechanism. I think the election commission itself had this giant conference where they presented some of these ideas. There are multiple speakers. One of the speakers during the conference was Mr. Jayesh Ranjan who is the IT secretary of Telangana who also said that even Telangana was trying to do something like this. This was a year ago or so. So the ideas are coming from everywhere except it kind of got implemented in Telangana first. Telangana has been trying to experiment several electronic or election-related technology-related experiments since 2014. We first did the Aadhar vote ready linkages that kind of created a huge number of people not being able to vote in 2018 elections. Then there was also these experiments of verifying voters using facial recognition. I think it was also kind of a dummy way where you are allowed to vote no matter whether it says yes or no, but they were trying to test it. This happened in Hyderabad in 2018 or 2019 elections, municipal elections again. This is the third significant experiment in Telangana when it comes to technology and voting. So we also kind of have become a laboratory when it comes to testing technology in India. So which is why you see Telangana doing it much before PCI proposed its own solution. So you will see something again from election commission of India later. So coming back to the elections in Kama, I think it was about a month ago. If I'm not wrong, October 7th is when the Telangana State Election Commission suddenly announced that they are going to host these small elections on October 20th and the voters can register on this particular smartphone until October 18th. So October 7th say October 18th, there were actually court holidays. That's our festival. So it was strategically announced so that nobody stops it in any way. Also it's a mock election, courts don't intervene in interfering real elections. So forget about this being stopped in any manner, but it was strategically announced. Anyway, after 10 days, after all the publicity of asking people to register for it, there were 3000 voters who registered for the mock election and 2000 or 2400 or some voters voted on the day of the election on 20th October. So on 20th October, I went to Kama. We kind of gathered a bunch of few people who registered themselves for this voting exercise. And then we sat and looked at doing certain things on influencing voters or the election, right? Like what all is possible? Since it's remote voting, we were able to gather a lot of people to a single place and make them vote for a particular party. Like I knew who he was voting for because I could see it. Even though I can't take a screenshot of whom someone else was voting for, I could still use my mobile phone to capture who he was voting for. There's an actual election happening in Telangana today, the Huzoorabad Bible. And there is active debate on how much money each party is paying for the Bible. And people are actually demanding that they be paid a lot of money for the elections. So this is important because in case of a remote vote or voting elections, I can simply call these people to my office or wherever I am or I can even go to their houses and make them vote for me, make them show me that they have voted for me and pay them the money. So what this is doing is that it's taking away secret ballot. So as Mr. Sunil has explained to you, one of the fundamental principles of elections is to have secret elections. You really don't want to know whom you're voting for. But at the same time, there is the other constraint, you want to make sure that the elections seem fair. It's a very conflating set of interests. You wanted secret, but you also wanted to be transparent because you wanted to be fair elections. And over ages between ballots and people throwing ink into the ballots and stealing ballots and bringing elections and getting EVMs, all of this process has been standardized, like even getting photo IDs. It took us a lot of time, right? Like it's not suddenly you change elections. It's a process in itself. Now, these new set of elections, I've told you about the issue of secret ballot that anybody and everybody can know whom I voted for because they can stand beside me because it's not a controlled environment. There is no polling booth, there is no polling officers, there are no observers, there are no polling agents. Every much takes away all of the polling setup. Now, let's come to the other issue of voter integrity and counting, right? Like after the vote is done, you need to count the votes. And when you actually count votes, in case of ballots, in case of EVMs, you still have observers. Party people are gathered to ensure that EVMs are being counted. The numbers are being reported fairly and there is no process of tampering. Now, in case of this entire election, the results weren't announced. So we don't know if Alpha was, Alpha won the election, Beta won the election, or Kamau won the election. It was a mock election, but still they did not announce any results. There were no observers during the election or during the counting, there was no counting. Now, they said once you vote, the vote essentially gets registered on a blockchain. Now, I really don't know where this blockchain was, like theoretically speaking, you could have brought a bunch of expert observers and you could have showed like, okay, this is how the registration is happening. And when you talk about blockchain and the issue of trust, whether you can trust your vote is being given to a particular party. Blockchain, the idea of using blockchain is to show that, okay, it's providing you integrity and trust because it's distributed and it's decentralized, there are so many of these people who are observing. We can say that this particular person has voted for this particular party. But because it's all the servers where with the election commission with no observers, we really don't know what this whole blockchain setup was. And it's also important to understand that the idea of blockchain itself came because people did not trust humans and they want to induce trust into machines or they designed blockchain for that where you can have mathematical algorithms verifying trust. Theoretically, it's a great idea, but nothing of that sort happened here. So we really don't know about this entire blockchain setup, even if there was something like that. So I will not talk about that aspect of it anymore. The other aspect I want to go back is verifiability, right, like also in the counting setup. So when I vote for Alpha, I really need to know that a vote has been registered to Alpha in case of a ballot paper. If I voted for the party, I know because I'm stamping on that party, I know that vote is really being registered for that party. In case of EVMs, there was this doubt. So they finally brought in BVPACs where they said, okay, whenever you vote for a particular party, there will be a printout of whom you are voting for that party. But then again, to ensure a secret ballot, the slip is not given to anyone else. The slip just remains at the police station. So in this case, even though you're voting through your smartphone, it might show you that you're voting for Alpha, but I really, really don't know to whom the vote is being registered to. So that end of the election commission side of blockchain and how the vote itself gets registered on the integrity of the vote and counting process. All of that is a black box. So I can't see it. There are no observers. I mean, I can't even see it from a machine. So there's no way I know what's happening on that side. I tried to kind of document some part of this on the API communications and all of it, but it really didn't happen. So I really was unable to gather a lot of data. So what I did is also to file some RTIs with the Telangana State Election Commission to provide answers to some of these questions on what happened. There were a lot of challenges when people were trying to register and they were trying to vote. Sometimes their other number and the phone number didn't match. It's facial recognition failures. The last time the election commission tried to experiment with the facial recognition system, there was a 66% success rate. So we really don't know how much the success rate was during this election. So and again, so there was no error manual. We really don't know if the election commission was trying to go document if there are any kind of errors that have taken place. But what they did was they had this email ID asking people to send any issues that they have to a particular email ID. So in all the election that happened, they announced it as success. I don't know on what parameters they have evaluated the selection because they were no observers. One can't really say if it's a successful election or a failure. So every election you ensure that you bring in observers, also international observers, the election commission of India does that for our general elections a lot. Like you have some observers, external observers in particular to ensure that nobody is influencing the election. So in all, all I can say is that this election was kind of an experiment with no result. They can claim that the results were great, but even we really don't know what they saw. So until we get more information. I will simply say that this election was just an exercise to gather some evidence and we really don't have any evidence at the stage. I think Srikanth is going to speak some more aspects on the code of the app now, like I spoke about what happened on the day in Kammam. Srikanth will come step in and he can explain to you what he was able to document looking at the code of the app. And then we can come back and we can discuss about some of this. Srikanth. Yeah. So I just share my presentation. Yeah, so firstly I'll put a disclaimer and say this is an armchair observation. So what I was able to document was what was all accessible from a browser about this mock election. So these are armchair observations and touch briefly around what archiving elections are about and what digital elections mean and what new opportunities we have. Some comments on this particular voting process and some of the app. So with that, I'll touch briefly on archiving elections. So the whole point of this talk is essentially for us to create material evidences for the future generations and this could be for a researcher 10, 20 years down the line trying to study the history of digital elections in India and then at a time when digital elections are complete and all and for that particular person trying to see material evidences. As we know that let's say anything on the web plan for 14 days. Roughly. So how do we create more data and materials that are available for a longer period of time. So with offline elections, what we've had particularly with the evolution of tech and the uptake of tech by the election commission itself. And a recent example for that is in India, I think post 2000s, the part of the electoral reform around election affidavits and candidate declarations having to be public. We have a lot of websites that we are available generally with the election commission and we kind of went a step further in 2019 with all the affidavits available are now available publicly of all the candidates who are standing in elections since 2019. The election uses this system called candidate management system, where all the candidates are supposed to upload their affidavits and all those affidavits are publicly visible. But, again, putting the challenge of web where a link last only for a few days, it's only a matter of time before election commission, maybe three, four years down the line revamps its website. The affidavits disappear. So what I done sometime back was to archive all those nomination papers. And we now have like somewhere on 50,000 nomination papers that have been filed since 2019 that are archived of the Internet that are archived. So, I mean, with that said, that's the only thing an armchair observer with a laptop can do in a real election, because the elections are offline. And if you want to kind of archive the election process in itself, one could kind of go through various newspaper reports, the results and data that is put forth by the election commission itself. And those are the data that one could kind of archive. But with digital elections, what we have besides archiving the data around elections, we can also archive the election process in itself. With EVMs, it's impossible to kind of archive, although in some court cases and so on, the EVMs are kept intact for a period of time until ultimately the case is decided. But with digital voting, like the app-based voting, it is possible to archive code. And so in that election, and so in that context, I kind of tried to archive whatever information that was publicly available about this mock election. So these include the apps themselves, the documentation, these include the circulars by the state election commission, the Google Play Store data, and some images around the mock election. And a reverse engineered source code of these apps and a list of media reports. So these are all available for anyone in the future to kind of have a look about this mock election. So as you mentioned about the missing blockchain, that's the same question that I also have. One might assume that the blockchain is probably available on the internet, ideally kind of entirely like forkable and anyone can join us alone. But we don't have any such thing available. So even I'm asking the question, where is the blockchain? And some data around this mock election, as we must test upon, I just kind of go into detail. There were three versions of the app that were released in the two weeks when the mock elections were announced. And here is the first challenge. And we kind of need to see, have more evidences and learn more around this part. One is like the data says, one of the media reports said there were 14,000 app installs between these two weeks period. And this was a media report by having a quote from the state election commission saying that these many people registered, downloaded the app, which meant that they were trying to register. But before we go into the fact of how many people registered and what registration means, one quick point here is that of these 14,000 app installs, only 12,000 odd were from Google Play. Now this poses a serious problem because how do you trust the channel in which the app is delivered. And this is important because you need to know that the app is actually original app that the election commission distributed for and not any other tampered version of the app. Although there is some amount of tampered proof checking in the app, we still need to kind of, the logistics of getting the app installed need to have a process which is again like fair and tampered proof and in a way where there are no dependencies. Again, one could say that Google plays an intermediary here and that could have its own issues. So that's on delivering the app. So this part is about getting the app to the voter. And now that 40,000 people installed the app, the next step for them is to actually register on the app. And we'll go through the registration process in a bit. But again, one of the media reports suggested that roughly around 3,800 people were able to successfully register, which meant that the remaining 10,000 broad folks dropped off. They could either be curious people wanting to try the app without having no intention to vote or eligibility to vote, or they could be people in that town downloading this app, but we're kind of unable to kind of get themselves registered. And I can see this as a parallel to any of us trying to register ourselves newly in a constituency we have migrated into and there may be issues of finding our name on the polling list. This is something of that sort where the registration in the app gives you the right to vote. The offline parallel could be getting a boot slip. But the only fact here is that we don't know basically how many people actually dropped off because of issues with the ticket itself. So the numbers are quite alarmingly high, but even if we count 50% of them were actually eligible voters who had downloaded the app wanting to participate in the mock election, that will be like 7,000 installs. Even in that they're like good 40% of them were not able to register because of some of the other issues. We'll see what these issues could be in the next slide. And among the people who were able to register, now these are people who are eligible to vote and who have a strong intention to vote because they have firstly downloaded this app and they were able to complete the registration process. So they are probably wanting to curiously vote in this mock election. But only 2,200 people were able to vote or those are the number of votes that were casted. This again could be mostly because of drop offs due to technology. I'm discounting, let's say, not voting as a mark in itself, unlike the real elections where let's say not voting could mean that you are out of town or you're sick and so on and so forth where you are not able to go physically and cast a vote. Those excuses are predominantly discounted here because those are precisely what are suggested as advantages in a digital election where it just takes 5 minutes for you to vote and you don't even need to go out and vote. So the convenience is there and these are people who are anyway interested and that's why they've kind of registered and spent some time with the app. So I would probably say that the number of people who got dropped off here are 99% dropped off because of a technology issue. And then the next thing is that the results were not announced in the election but a couple of things were kind of documented. One news report referred to the fact that there is a dashboard in which selected users have access to live results and this could have issues around having access or knowledge of the current voting behavior and this precisely was happening in Estonia where the results are actually live. In India we don't, we conduct the polling and polling is closed and results are counting starts two days after that even in a multi-phase polling scenario the results are kept withheld before the other phases of polling is complete to kind of see that the influence, I mean the results do not influence the voter whereas in this case there is a possibility of live results getting collected especially irrespective of the fact whether that is kind of accessible to political parties or election officials the very fact that these, I mean the, there's no separating as counting and the results are kind of streaming in to and is available for anyone having access is something to worry about and there are issues around this live results in Estonia as well they've been documented as part of their digital elections. So the next thing is this is from a document that the election state election commission puts out this is about the registration. So the registration window in this case was open for 10 days so what does an average voter need for registering they need to be having their voter ID and mobile number with OTP detection the app takes this permission but what is also needed is that this should be the same mobile number to which your ADAR is linked so this should be your ADAR link mobile number and this is because the first step it does an ADAR OTP based EKYC and this EKYC then downloads an offline ADAR or an MADAR file on your device and this is then further shared with the election server. So this is the first step the first step is of course entering your voter ID and getting an OTP from the election commission the second step is performing in ADAR EKYC which again requires an OTP so these are basically two OTPs to get to get past this code and the third is voter verification where your ADAR name and your voter ID name are matched and they say 50% name matching algorithm. I mean I really don't know we kind of need to know this name matching algorithm because a lot of the exclusions that were happening because of ADAR be it in the LPG system the LPG system used name based deduplication or the Thimstinwas has a case in the Telangana High Court on the voter relations which also happened in Telangana, under population Telangana both the states by having an ADAR EPIC linking and that also kind of dropped off many people from voting and we are kind of seeing the same algorithm getting reused here in the digital loading except for the fact that this is one step of the process in digital loading but this step has been there in place at least in the Andhra Pradesh and Telangana where there is already an voter ID ADAR based cleanup which has already happened but this is the contentious provision of voter ID and ADAR linking being in place but this is also a part of the step in the app to verify the order whether that this person is the same person who is voting and who is also registered with the same ADAR and the name matches and only then you go into the photo verification and the photo verification is a liveliness detection selfie that happens and this photo is actually matched with the EPIC photo and here it's important to note that this is not the ADAR facial data but the facial data that is held by the State Election Commission as part of the voter votes and your face is matched against that and once you're done your registration is complete so you need to perform all these four steps to get your vote slipped to verify that you are the person who is there in the voter role and this is your device and once you do this basically your device is bound to your voter ID and your digital vote can be cast fully from this particular mobile device now if you happen to drop your mobile device and your screen gets damaged you cannot vote from any other device and when the registration is closed the voting is activated on the polling day and from what the title says is that we may not actually have a fixed voting day if this were to happen and this is kind of pretty much similar with the US style of elections where there is no single polling day and it's a running election so when the voting module is active this time you enter your voter ID and your mobile number and IMEI is automatically checked so this is to confirm that you are voting from the same mobile so this is basically a device binding check in technical terms which also happens with a lot of digital payment apps and so once this device binding check is complete you are again asked to do a facial recognition and once this is done you are allowed to vote and then the voting screen comes and you are allowed to vote and this is what Kotali just showed us in the video so what are the authentication factors one is device binding which binds your vote to your device and this is done after three stages of check one is an OTP check by the election commission the second one is an ADAR OTP based EKYC and the third is facial recognition with the voter votes so these are the three checks that are there to grant you that vote so once you are there you are registered a point to note here is device binding itself has few issues and we have loads of experience with UPI simple thing like changing the slim slot or in some cases there are issues with certain telephony providers capable of having device binding successful because of the technology they use and these are limitations in UPI which are loaned in case of device binding and this will get carried over and the other prominent thing that kind of got noted was OTP failures but we don't know whether they were specifically UIDI, EKYC failures or the Telangana state election commission OTP delivery but OTP delivery failures were again a good chunk and lot of people were not able to register and this probably explains the drop from 14,000 installs to 3,800 people who are able to register and as for the process one device allows two voters to register so maybe like in a family of four two people have a vote, they can vote using the same device and I think I covered the excessive drops in the registration phase is more likely to the fact that there is an other EKYC and there is a name matching step that is there besides the facial recognition the facial recognition is there in both the steps for registration as well as voting and there is no other verification during the voting phase and coming to the facial recognition part of it so Telangana uses their this election, Telangana has its own real time digital authentication of identity, this is basically one could actually say it's a parallel to other except the fact that here you don't have your own other number it's only the state that has your ID but the state can authenticate you against this particular system and this is basically a facial recognition system which Telangana state has kind of licensed from a technology provider based in Geneva called Tech5 and it was using this system that the 2020 pilots which the universe was mentioning that happened in Kompalli municipal elections just in the outskirts of Hyderabad and there were like 150 polling booths on which the facial recognition were tried again we don't have any follow up reports or documentation around the success or failure rates of this pilot but there were facial recognition pilots that were happening in the 2020 elections and I have documented the election commission circulars around that effect and this of course was one of the SMS confirmation that one person had shared on the internet which after voting they get a successful message and what we were also able to verify was that the voting is allowed only after kind of checking the update to the app so the app needs to be updated before voting and what is the app tells the app tells there was very minimal information around what the app does because the app did an UIDI offline other XML based VKIC it had some tamper proof version checks though this was like using 3-4 year old code to do the same and I don't know the effectiveness of this and the app loads a web view that is powered by React Native for all the functions it does so in that sense like we don't know like what the app does because there is very little code around the app all it does is to load the web view that's inside the app and if there is any other information that you have come across please send a full request to this repository my intention in documenting this is for anybody in future to kind of have a look at and evaluate how the selection went this has to be like one repository of information that's available for anybody to kind of add to their work with that I will conclude my presentation thank you Shikant for that detailed description and also thanks to Srinivas for his insights that he shared now let's see if there are any questions do I see any questions in the chat so while waiting for this to happen for the questions to appear you can have a little discussion I have some questions of my own can we see you also Srinivas I turned on my video you can't see me no I have turned on my video I can see yeah hello I only seem to see Shikant and no one else but that's all right so long as I know you are there so some of the thoughts that came to my mind while we get the questions I would ask either of you to respond and you know the first was that some features of the app are like the proctoring software that's now used to ensure that students can write their exams from anywhere from their homes neither the students nor the teachers seem to like that software and there are plenty of issues and it seems to me that the same kind of issues we might have will be voting some of them not all the issues will be similar and it may be worthwhile looking at those apps and reactions to those apps to know how this might perform in the in the world do you think there is something to look around there I'll just add one thought here yes but I think this app does far little less especially around the concept of proctoring because proctoring apps kind of deeply know your device and this is kind of a pre-requirement there because proctoring app should ensure that you are not using something else to cheat in the exam and so on and so forth I think this app doesn't have those features precisely to kind of give that privacy and this is both blessing and disguise because if you kind of have that then it's kind of more rigorous and we've kind of had this issue around not just on this context but even there was a PDM once did something like this to kind of disallow the screen sharing apps because too much fraud was happening and people are suddenly like why should PDM tell me which apps I should have or not on my device and PDM did was in a pressure to do so because a large function a large section of their clientele were actually susceptible to the frauds and hence kind of decided a tech solution to kind of detect which apps does the device have whereas I mean that came to attack from a different section of users who wanted complete autonomy in fact they did not want PDM knowing which apps are getting run and so but again in a diverse country like India that's that will always be a problem where like one section would want no monitoring the other section would like monitoring and monitoring in this context could also mean that how do you ensure that Pegasus is not installed on my phone if it's the duty of election commission to ensure secrecy of the ballot now how does this app detect that no other app is capturing screen recording and I can say from the work we have done around digital lending where screen recording is pretty much a native feature in all those digital lending apps and this is for a reason why those kind of apps have those privileges which is to kind of have it's a different debate on digital lending but there are apps which actually have screen recording capabilities and background capabilities which might be installed on the person's phone now is it the responsibility of the election commission to have detection of such apps in this particular app to ensure secrecy of the ballot or again how would they do that without kind of peeping into what all apps I have because again that's not the piece of information I want to give to the election commission in real world I don't give any information when I go to vote except for what's minimally required which is my voter ID but how do it's technically a challenge for a kind of election commission to ensure that my device is surveillance free and Srinivas was mentioning that it's possible for somebody to actually send people to home if you are vote you don't need to do all that you just kind of go and buy from a vendor like Pegasus and you know who is voting where and that's far more scalable and how does let's say a digital election counter that I think that's something to be discussed yeah I mean more than Pegasus I'm afraid about Google like it could be Apple like now the amount of data that's being sent to some of these organizations you really can never claim that it's a secret election unless the election commission controls the entire device which is why in case of EVMs they consistently insist look we manufacture the entire process only we know how it functions nobody else does it there's nothing going out of this box or coming into this box because if you look at the kind of claims that happen around that one can hack EVMs with radio waves okay one can tamper with it which is why access to EVM is so highly debated in India whenever you see EVMs being taken by non-election commission officials so this idea of integrity of the device itself very important when it comes to electronic voting whether it's remote or not we can still have a smartphone based voting a blockchain based voting but you still need control environments yeah correct yeah thanks so there is a question from Vandana.org she wants to know that she doesn't have a Hyderabad address on her radar card she can register for a digital process in future I don't know whether she is asking about it in future or whether the next election think of this scenario like you're talking about remote voting you're talking about voting for migrants you're talking about voting for NRIs now this concept of space and constituency is changing okay now when it comes to ADAR one can simply change their address like we can just go online to submit ID proof and change it now in a country when elections are held consistently across the country yeah with remote voting options now how easy or hard it is it's both an attack on elections where you can get people from a different state to come vote for you or you it could also be an issue for people in the locality when they are unable to change their address formally digital processes so again no election will be 100% remote voting like even in Estonia when they did this they had partial remote voting you still had boots it's not like election boots went away you can't replace it because you will always have population which can't access these devices and you can never say that there will only be remote voting unless you are giving everybody a phone even then there is also challenges of literacy so in terms of this very particular question on constituencies and spaces and delimitation the idea of remote voting if you are bringing in changes election so drastically the representation itself needs to be re-looked into let me ask you a very different question both you and Shrikanth got the issue of convenience in using the app at lunch anytime you can vote no need to stand in the queue no problems but let's look at the converse standing in the queue for some people who otherwise live very productive sheltered lives cars and air conditioned offices and homes standing in a queue is a rare opportunity to stand with your fellow citizens your vote is going to affect their lives too and everything is not about convenience after all people stand in a queue to play at a temple why don't they do so quickly endlessly from their mobile phones whenever they may be this is obviously not a technology question but then we are not talking about technology alone we are talking about how people as human beings think and feel and from there they decide to vote I don't know whether everybody has a clearly formed idea of who they are going to vote for to be reached for polling station or what percentage don't have a clearly formed idea or are still thinking so do you think those kinds of considerations also come into play shall I go first yeah so I think it's an important point that political scientists have been trying to understand over ages if you compare rural polling percentages versus urban polling percentage you always see that the urban polling percentages are low and especially when it comes to municipal elections this small election was held for municipal election voters and also this was kind of targeted to a particular people right like people who are educated people who probably have a smart phone and you look at the numbers so when you look at existing elections in India and how urban voters don't seem to be interested in municipal elections even compared to the general elections municipal elections get further low like it was 30% in Hyderabad JMC elections in 2020 when the election happened so can a smart phone based elections or remote based elections if you make it as easy as possible can it increase this and the reason I say this is that what drives rural population to come out no matter what and vote is not what drives urban voters their needs what they want from their representatives like they want certain things which the urban population already has so I don't think making elections easier is simply going to increase the turnout it's what to do with politics and what people want and which is why actually in case of Hyderabad because there were large scale urban floods during a municipal election just before the municipal elections there was a massive increase in turnout so sometimes there are these causalities which could drive things but making the elections easier just because there is low turnout doesn't really make any sense yeah so okay there is a question from Mansi she hasn't put the question but she has made a request so maybe she wants to ask it in person Mansi yeah please go ahead I just thought it might be better to ask so actually this was with respect to one point which was made about how it might be possible to see who you have voted for because somebody could come to your place and check who you are actually voting for that actually got me to thinking that I mean of course it is possible however it might be difficult for them to send their workers to every household to check who is everyone voting for but that got me thinking about the very important role that all these political party workers and election agents etc play at the voting boot so they are there to also check that nobody is doing any electoral malpractice and I know and we have also come across a lot of these instances where people sometimes report that the EVM is faulty and it is unless and it is only when some of the political party agent who raises the voice about it they might not actually change the faulty EVM machine etc so I was just thinking about the fact that these and I have also worked with the political parties and I know that these are people who are trained and who have been observing elections over the years and they play a very crucial role in ensuring free and fair elections that happen at the voting booth so I was just wondering about the fact that they might also become redundant in the process and the role that they are playing which I think is a very crucial role will also sort of there will be no one to actually find out for instance if the app malfunctions if it has been rigged and if whichever button you press the vote goes to a particular political party only and then you really have no recourse there so yeah this is just an observation not really a question but yeah just wanted to share share this point so things change once when technology is introduced some changes will happen some things will become important some things will become less important observers will always be important because unless somebody is observing how do you know what the hell is happening and they have to be given means to make those observations Srinivas brought up the issue of verifiability and where is this data on the blockchain who can read it, how can they read it and unlinkability that is the data on the blockchain cannot be traced back to a person all those issues are there which are issues of design and I suppose unless these details are made available nobody really knows we have very little information right now on on what was behind that so that is my two bits that Srikant and Srinivas are welcome to give a more detailed answer I will just add one bit to the convenience argument but I think we are talking about convenience of the order we should also talk about convenience of the election commission itself and the state apparatus I think it is convenient for them to do digital that is what they think in the sense that because I think the current system relies on the use of extensive human resources we are talking about literally moving hundreds and thousands of people across from different places and kind of have that set upon a single day and the kind of effort that goes election commission actually feels convenient to have an app based on and this also has an economic implication where there is a cost of election that is there today and this cost of election is drastically reduced and one could even say that I mean this is the same argument around cash and digital payments where instead of moving cash across if you are able to have a digital payment system which is far more efficient and that argument is being repeatedly put on and that is the whole premise on to which let's say DBT or digital payments for payments for everything kind of is prioritized over cash because cash is considered to be costly it is resource intensive to move around and it is cumbersome and so on and digital payments is far more convenient while on one hand it is convenient for let's say we could do e-commerce we could pay somebody else remotely instantly and so on and so forth the same parallel arguments can be seen in the elections it is very convenient for the institution that is conducting the elections it is a valid one it is very convenient for the more convenient for the institutions more economically cost effective as well I mean one could actually say that we can even give you a phone and you could vote instead of kind of managing the election in a physical way and this will probably kind of set off in two elections or something like that where the costs are offset that is the same argument it says that we can even kind of subsidize mobile phone which is their currently part in some states where there are free mobile phone schemes and so on so there is a I want to add that part of state institutions specially election commissions trying to adopt blockchain like this is not a very Telangana state election issue anymore because we know IIT Madras, the Tamil Nadu government agency and the election commission of India vastly interested so where is this interested blockchain coming from I am not really sure see bureaucrats understand the idea of blockchain because the idea of blockchain comes from lack of trust right like the people who built it, people who built bitcoin were crypto anarchists who believe that the power structures that held that whole governance are not fair and that you can essentially ensure these systems are fair by ensuring mathematical proofs and the way bitcoin and blockchain theoretically and in practice the way it works ensures this in certain scenarios but I don't think the way they are implementing it is not how crypto anarchists itself have probably pushed for right like they think it's giving them trust but we don't blindly trust the election commission of India and that trust has been going down forever like if you look if you go ask people political parties do they trust election commission of India they probably did sometime back they don't now anymore maybe in future if they had someone like the institution back oh yes I will trust the election commission because he is leading them so this idea of how trust work functions trust in technology trust in who is leading the election commission all of these play a factor in elections they do let me put it like this suppose the blockchain was used in such a way that the election observers that fancy was mentioning could verify and satisfy themselves that every vote has been cast correctly and properly without being able to trace back to that voter of course and then they sign every vote when it is cast immediately after it is cast that yes it is correct this may not happen at your home it may happen in a polling station again but an environment could be created where the people remain and now they are able to count the vote themselves so that actually increases the trust because the election commission provides a way and anybody and everybody can verify that yes elections have been conducted the same way that still doesn't take away the problem of providing you know I mean protecting other kinds of harms that you have mentioned but at least it will be an enhancement over physically transporting the ballot papers or counting and or doing counting in a manual manner in some corners you could end the day and then you can count everything and anybody could count if the blockchain is distributed across political parties and for independent observers yes true possible so it's actually a question of what we wish to do with policy and how we implement it and we know so little about it right now so that expects to ask how it is proposed to be used the architecture the code all of these we know nothing about it and this is something that we consistently keep bashing the government to like give us code show us what you're doing right never happened we did have a small win elections it's really important which is why people keep insisting EVM's source code show us how the EVM's work so until that happens there's this lack of trust so there's a question from Sankarshi she wants to talk he wants to ask a question to you and to Shekha so Sankarshi go ahead please oh yeah I want I think I have a lot of questions as to all of us because we know so little we are concerned about so much and the fact that there is a lot of ground to cover but I have a very simple question in terms of voter suppression in elections has been a particular area of concern over a number of election cycles and based on what we see as technology getting introduced especially this one do you see it being causing more of this and being the catalyst for more voter suppression more targeted voter suppression and it makes it easier for anybody in the world as the power of the state do you see this as something that we should also talk about among the various harms that we discuss so maybe both the speakers today will have things to say I am sure okay so I think there were a couple of instance where there were errors errors with no outside play that was designed into the system voter peace not working racial recognition not working internet not working so all of these depend the election itself the way voting itself has depended on so many of these outliers and I don't know how much the state election commission gathered but in any other scenario yes you can hope in point where these problems are all it probably takes is to take out a cell phone tower in a particular village to ensure that entire village doesn't vote for a particular political party you could potentially take out such diesel generator you can just blow it up with fire farmers have done this to GOO stations and I am talking about political party scale now a nation state attack could be far more sophisticated where you could have a malware kind of scenario which is not other being only a select participants to not vote and even political parties could hire groups to even get this done if we go to such scenarios right the election commission itself may not observe this but individuals who are affected by it will be yeah so what you are saying is that elections can be the individual I mean their answers can be attacked or for the entire infrastructure itself could be could be attacked and then we lose it to people outside the country I mean it's not the political parties it's not the election commission it's not the people of India but somebody else can also interest here in there quite possible because unless you are saying you will build a great wall of India's internet and or you are saying that all of these will be through an internet where election commission is laying its own fiber of the cables to say that look network won't be disturbed because we own the network but are they going too far actually assuming all of this because the initial announcement only said that this will be a substitute for postal ballots hardly very few people do it like that so if they are given some means maybe it doesn't affect really the functions or the way it was pushed in Estonia and the way it functions it might come on postal ballots so I would like to kind of place it on fact that in 2016 a DMK candidate lost by 160 votes and this was primarily on the postal ballots and the case went to Supreme Court and the Supreme Court took 5 years to decide the case and he is currently the speaker of Tamil Nadu he got elected in 2021 but his 2016 result was actually challenged for full term of 5 years and he never became that MLE and so and just on the limited point of this is only for postal ballots or this is only for 150 votes or very small percentage of votes which is kind of on postal ballots we already have an example of us having issues in adjudicating despite all the let's say election in Supreme Court rulings on election cases to be decided in 90 days or what not this was the Supreme Court in itself which took 5 years to decide on a case which had to do about counting of 160 postal votes now combine it with the digital illiteracy of the highest courts which they are certainly in their own capacity there is a lot for the courts to actually improve themselves when it comes to adjudicating technology and if the courts could not do settle a case of counting 160 postal votes in 5 years I really don't think we are kind of safe even with that kind of exception where this is only for a small percentage of votes because in a cutthroat election these do matters yeah I mean if they won't matter then why give them the right to vote anyway because they don't matter but if you give them the right to vote it won't matter it must be included correct well but to be fair or rather to to be fair to the courts they don't have to really adjudicate on technology matters and the law and the individual application of law and cases that would be great there are other people people like you who can look at technology and who can point out what it can do or it cannot do and its consequences are things that other people can also understand very well and they can choose to trust one person or the other and you know do their job any other questions to AC are there any questions by chance on no we haven't had questions on YouTube yet I think I mean yeah to kind of push this conversation along one of the I think the challenge that I often see that these engaging conversations and discussions on technology is that those who are very proficient in the technology tend to get down into the weeds and dissect the appropriateness of the technology but this actually the technology part of component of this comes quite late into the game is what I feel is so we probably would also want to have some ideas in terms of okay now there has been a mock detection that has organized seems like it has been very small in terms of the footprint with not enough clear declaration of intent and goals and hence no clear evaluation of how it was a success and Trini was mentioned that he has already has some RTIs in the pipeline that has been fired but what else could one do and what are the things that one should look at I asked this also in the perspective of this report that came out yesterday I haven't read it completely that in Bihar there is optical character recognition being used to identify the vote voting and then there's an artificial intelligence system I'm not really sure how OCR and AI are now getting munched together but anyway let's just quote from what I read this on the media and I'm a portal that came out late yesterday so I see there is a whole bunch of new fangal deep tech or complex tech being introduced into processes that have been functioning for a while and it is becoming more and more opaque in terms of the systems design systems requirements evaluation criteria architecture and as university code to be made available so with those kinds of opacity in front of us what should anyone who is invested in this topic how should they proceed what should they be doing what should be their main areas of concern I think anybody who is invested in this and it's not only citizens who are invested in this the state machinery industry everybody all all organs of the state are invested in this I think the only way is to remove the opacity and to be open we heard that the experiment we don't know the results of the experiment because we don't know the results of the election the mock election that was had I would say a successful experiment is one which answers the question so you should know which question were you asking when the experiment was made a good experiment is one that throws up new information new ways of failure new modes by which something could go wrong that's also a good experiment because you can think about it so the technology comes and embeds itself everywhere that's alright we need to have I think a very open conversation completely open conversation all actors all participants everyone has to be open about it all code everything has to be in the open or everything has to be possible to see and then alone can better systems merge and appropriate use of technology can be made going back to blockchain and bitcoin which which we must raise the originators of bitcoin didn't put up hardware or all the massive infrastructure that is now running they wrote a paper they published a protocol to start with isn't it they didn't hide in you know how it works you can do mining yourself you can join or withdraw from the network as you please so systems are good if systems design can be out in the open to be examined the code is good if it's open for examination and test and can be can be improved if people are actually invited to improve the code if somebody is good and he can point out an obvious flaw he can possibly also contribute to improving it so I think openness is most important in discussions in considering all ideas and just pointing it out to everyone that is not giving the information is not on that's not possible that's open to legal challenge and all kinds of challenges and therefore and I don't see anybody would be interested in hiding anything or should be interested in hiding anything even if it's wrong even if your software doesn't work today as you expected to okay just share it that this is what it is these are the results this is how it happens what is the alternative what people will suggest and if they can't suggest then they will conclude that the technology cannot be adopted yet for the purpose of which we wish to adopt and some other day we will be able to do we can revisit or the same technology can be put to use in a different manner right so I think encouraging activities participation and being very open is the most important thing in this and of course you yourself have been looking at it maybe you could share your thoughts on what others could chip in I'll just chip in on the openness part and there are challenges in that and I'll kind of bring in a specific example in this case itself right where facial recognition was the other tech function that is distinct in this election besides the blockchain now assuming that the blockchain is an election specific infrastructure the facial recognition part of it is actually not and they've kind of deployed a proprietary solution here and today these proprietary solutions have a market to serve and it is not probably in their best interest although there are open implementations of let's say efficient recognition as well we don't really know why let's say there are open implementations it's kind of fair to assume that the proprietary implementations are actually kind of having an edge which is what at least the facial recognition technology right now it is so that being the case there will obviously be kind of resistance by the technology providers themselves to be kind of be open because that kind of takes away their entire business so this is one kind of challenge especially when we are dealing with newer tech where there are limited open options and in fact if you want to actually have this is like using Linux for desktop in 1995 while today one could kind of reasonably use an open source software fully to perform all the functions in a computer back in 1995 if you are a bank and you had to have an IT system there was no choice but to have so this is kind of we are kind of in that phase and in that phase where there is not much development around the open infrastructure of let's say the underlying components facial recognition being one here there will obviously be issues in taking up such technologies and making them open yeah but the simple thing is that if you are using a proprietary technology at least it should be open to test the results should be known success rate type of faces it can recognize it cannot recognize biases that are known to exist and all that kind of testing has to be done you can do black box testing for something that you are not allowed to open source but then you know the limitations with which you are working I guess these are the kind of questions that can be industry debated and is good to endlessly debate them it is just not good to debate them all in one session so so maybe since we are coming the end of the time that was a rocket for this discussion I would like to request for the speakers first to to kind of kind of give their assessment of what we just discovered or what we think are the important points and then we move towards the close of the session if that's okay yeah okay so I think it's really really important for us to look at what's happening with the election infrastructure especially the election commission's interest linking other mind you they have sent a representation to the law commission of India and the law ministry to bring a law to do this and that's not about just purification of voters it's more to do a voting the way it has taken place in Estonia it's coming from there there are certain interests which are pushing for it and you're right when you say it could be good intentions or bad but these are the parties that are pushing it and those interests are not being made clear to us when it's being proposed to be passed through the parliament where you bring a law to clean up voter ID and other and here in this mock election even though this was a mock election we actually had real other data and voter data linked even though a supreme court order already exists it says you can't do this it shut it down and we can't just do this the way back as 2050 so these experiments however they are trying to do they're not making their intentions clear they don't probably have to but as someone who has been documenting this and who sees dangers associated with this we will continue to want why it's so important for us to look at these in detail however small they are the plans are being moved has nothing to do with postal ballots but to actually replace existing systems. So I think we need to be really careful about how to allow these systems even in ideal scenarios where I know we've discussed some of these ideal scenarios but without actually debating this and I think there's a way forward. It's not like I will keep opposing them and they'll keep pushing it. We need to have, we need to rethink how to redesign these systems and consistently we've been trying to do this outside the government ecosystems, outside the think tanks we're proposing these systems. If we talk to other people in the academy like people like Professor Suresh Benerchi who have been telling us how we can design these systems and that fair debate is not happening and the only way to bring that is to actually mobilize people to educate them and I think that's a process we need to do continuously and that's potentially happening now. Let's hope that more people get to understand what eventually I'll just end it there. Yeah, thank you, thanks. Yeah, I'll just place thoughts around openness because we've kind of had this rough ride on the EVMs both the initial ones and then the EV pads and they're still not done with it. We still have I mean every time around an election is done we have this issue around like how was the EVM tampered or was it tampered only one booth and that affected everything else and why is VB Pat not countered and so on and so forth. We have a running issues with the existing system particularly around the trust and one piece of place where like this trust could be enhanced is basically increased openness while that is not possible in a physical setup, in a digital setup it's slightly more easier to be more transparent and that's an opportunity. Now without going into this election or let's say how where the blockchain was or whatever in general a digital election can actually facilitate much more transparent verifiable archivable elections which are completely privacy friendly as well and this is a possibility and there can be experiments that can be conducted in in such a manner where these are experimented over a long period of time and certainly the the goal should be kind of having all the trust check boxes take because currently it's another debate whether what's the form of elections the form of representation whether we should have proportional representation and so on and so forth those are different debates and we are currently kind of narrowly pointing around the mode of voting and we are not even going into places of the current system of representation which which we have and so on on which there are issues again. So on this narrow point one key consistent thing that has been around is on the openness and this could actually be used as an advantage because you could say that look we were not able to kind of disclose in an EVM setup which we are now able to kind of have and fully open code a blockchain that is open and inspectable observable can be built and that's an opportunity if the election commission wants to use that as an opportunity and kind of make it make the lives of everyone easier because why would anybody want to say I don't have trust in the election because if if they can actually have it. Yeah, fair point, fair point. You said it absolutely correctly that why should we not improve the trust rather than say that you know that that we don't trust. So great it was very interesting listening to all the thoughts yours and Shrinivas as well. I think we have no more questions so we can end this session for which I will hand it back to Sankarshi. Sankarshi over to you. Well thank you. I would like to begin by thanking you Shrinivas and Srikanth for making time. I think this is going to be the first of a few conversations we will have. So it's not like this is the only one that we are having and then we are not going to have it because the topic is not just about e-voting, remote voting, but it is about voting, elections, democracy and the whole lot of other topics that are linked to democracy. So it intersects with those who are advocating for rights, inclusion, privacy, agency and there's a whole bunch of them and we haven't even gotten to what has been a pet peeve of mine that all these electoral processes that depend on mobile voting or voting through smart devices don't seem to also talk about accessibility when they talk about inclusion. So that's another topic that I would like to explore as part of these when we get to it. So thank you so much for making time today. It's been a very very engaging conversation. I made some notes this session obviously is getting recorded. We will publish the notes, the link to the session. I posted a link to our telegram group of Karana where a lot of these conversations happen. So those who are on the session today if you do use telegram and you want to participate you can choose to join that group. We are also pulling together as Shantar posted on the Zoom chat a two-part conference the first part being in December and then the second part coming up in January on this topic. So we'll keep you posted on that and how things go. So I think that's pretty much it. We will be providing updates on some of the topics that we want to discuss through November and December and if you are investigating working, researching on some on any topic that you feel is relevant and intersects with this area that we are examining, reach out to us through the telegram group or through the Hasgate page and we can certainly start up a conversation and see how we can have another public discussion and deliberation set up. With that, I have a nice evening if you are joining in whenever it's evening for you and those you will be watching to the recording in the future. I hope you enjoyed this session and that's about it I guess. Thank you so much. Thank you Shantarashan and thank you everybody who joined and participated. Thanks a lot. Thank you. Bye. Just to clarify, I don't stop the recording right?