Media reports1,2 about the prototyping of blockchain based technologies for elections in India provide a unique opportunity to examine and discuss technology interventions. Elections in India have not yet been heavily dependent on technological systems. It took four years3 to transition from paper-based systems to Electronic Voting Machines (EVMs) and twelve years thereafter to EVMs with Voter Verifiable Paper Audit Trail (VVPAT). This interim report is an outcome of discussions, webinars and conversations stimulated by One Vote Project.
One Vote project was initiated to explore the introduction of technology in electoral processes. ElectionTech and was initially hosted by the Kaarana collective. With the extensive set of conversations and discussions, it was evident that the project needed to be a defined with a space and forum of its own. The project is supported by Hasgeek
The current report is a summary of the conversations that have taken place - via master classes - since May 2021. We have also included some forward looking statements which provide further opportunity to investigate issues arising around the topic of technology in elections.
Elections in India are often described as the ‘dance of democracy’4 and the media gaze is relentless on the days of the polls. This has led to a situation wherein the following statements often ring true;
- The awareness about polling and related ceremonies is high, but the knowledge about elections is low. This dissonance is particularly visible in the keen knowledge of electoral rolls but not of the process which records legitimate votes as cast and available for tallying. This warrants public outreach about the technological aspect of elections.
- Explainability of technology is an important safeguard.This does not entail that the intended consumers of the technology must understand its workings in deep detail, but that it must be possible to communicate in a reasonably simple manner, the aspects of the rationale, the fairness, the safety and performance issues around the technology. This form of meaningful transparency is the key to building trust.
- Identity, identifiers and voting rights are intimately related. This is especially so when the presence or absence of one’s name from the electoral rolls is based on contentious rules and decisions. The electorate is often keenly conscious of the significance of being ‘othered’ or ‘marginalized’ and hence is engaged in establishing a verifiable and immutable record on the rolls.
- Political parties, civil society activists and technologists have a responsibility to act now in order to create a broadly understood framework around responsible technology that can be used in public services.
Blockchains are a type of type of technology implementation which belong to a class of software designs more accurately described as Distributed Ledger Technology (DLTs). Since 2018 there has been a growth in the interest and hype5 around DLTs and especially blockchain. It has often been proposed as a solution to an ever expanding array of problem domains6. Blockchain evangelists have been known to propose blockchain for all sorts of issues7 which come up e.g. financial technology (FinTech); provisioning persistent digital identification systems for refugees; complex problems in supply chains and in general any business transaction which needs to be disintermediated.
The over promotion of the technology has resulted in a situation where a perverse motivation stemming from the ‘fear of missing out’ (FOMO) leads to the absence of any reasonable assessment around the sufficiency offered by existing technologies and software architecture. There are situations wherein the specific advantages of blockchain technology or DLT are not evaluated with respect to the desired outcomes.
This report highlights some of the key aspects which are important to consider when engaging in any discussion around the introduction of novel and emerging technology design for large-scale publicly available services:
- Awareness around blockchain and app-based voting exists but knowledge about the same is scarce. This is specifically significant in context of understanding the constraints and limitations imposed by such a choice. There is thus a need for better consultation around the pilot projects being undertaken and the plans being made for future deployments.
- Commitment to a deliberative process as part of public decision making is necessary to establish better understanding of the proposed changes in terms of benefits and societal impact. This is likely to lead to meaningful transparency among the stakeholders and participants in such design approaches
- The adoption of approaches such as the Gunning Principles of Consultation8 are likely to address the absence of lack of access to complex and contextual information which results in being unable to complete a scientific evaluation of the proposal.
- Governments as well as private organizations contracted by the state must begin to include a societal impact review of Public Interest Software such as voting technologies. A better understanding of the readiness and completeness of such software is required prior to placing the trust of the citizens and voters on such digital workflows.
The interim report from One Vote is not designed to be a formal record of proceedings. Rather, it is an opportunity for the reader, and especially the lay reader, to familiarize themselves with the key issues discussed so far. Additionally, it attempts to put together the initial aspects of a model - a pattern - which can be used to examine the introduction and intervention of technology in the domain of public interactions.
While the sections, which are presented as individual chapters, can be read independently, we recommend that you read this report as you have received it - in sequence. Each section is a separate submission, but we have ordered it according to how we would like you to read it. The text of this report is written to provide a narrative sequence which explores the history of the electoral process in India and builds up a framework through which to evaluate the risks of technology interventions bereft due process of accountability and transparency.
We would like to thank:
- Ornit Shani, Nanjala Nyabola, Tarangini Sriraman, Subhashis Banerjee, Srinivas Kodali and Srikanth Lakshmanan for conducting masterclasses. A substantial part of our understanding around this topic came from their sessions, and the papers published by Ronald Rivest, Nic Cheeseman and Ben Adida.
- Taha Ali for reviewing the report.
- Nadika Nadja for helping with the structuring.
- Zainab Bawa for overall editing and encouragement.
Sankarshan Mukhopadhyay works on digital identifier technologies and examining how digital public services are introduced for citizens.
Chantal D’Costa is Research Assistant at the One Vote project. She is a Humanities undergrad at Azim Premji University (APU).
Anish TP illustrated the report.
PTI. E-voting to become reality soon? EC working with IIT-M on Blockchain Technology. https://www.livemint.com/elections/assembly-elections/evoting-to-become-realitysoon-ec-working-with-iit-m-on-blockchain-technology-11616755074358.html, 2021a. [Online March 26, 2021]. ↩︎
India’s electoral democracy: How EVMs curb electoral fraud https://www.brookings.edu/blog/up-front/2019/04/05/indias-electoral-democracy-how-evms-curb-electoral-fraud/ ↩︎
One Vote Report Annexure
As part of this project we have been hosting webinars and conversations to explore and discuss some of the key concepts related to better understanding of the elections. The complete set of talks including links to the videos are available at the Election Tech project.
This is a summary of the key concepts and highlights from the session “Voters Before Citizens” by Prof. Ornit Shani.
This talk was organized as part of the project examining the introduction of technology, and particularly blockchain as part of the electoral process in India.
Prof. Shani built her talk around the preparation of the first electoral rolls on the basis of adult franchise (in readiness for the first election) and how in doing so a commitment to the principle of inclusivity enabled taking bold decisions and a way of shared learning. A fascinating aspect of this process is how the decision around complete adult franchise was agreed upon and steps were taken to implement it even before the electorate became citizens of India. The preparatory work for this started even before the enactment of the constitution. Shani discussed what we could learn from that process about the steps taken to build a robust electoral system.
An electoral roll forming the basis of universal franchise, prepared and maintained accurately, was the plinth upon which the institution of electoral democracy would rest. The history of this process would help understand the values that underpin democracy and the electoral system. And while often some of those fundamental values are taken for granted, any discussion around them is needed to frame a critique of changes being introduced in the name of more effective democracy.
Prof.Shani has extensively documented in her book, How India Became Democratic, the work on the preparation of the first draft of electoral rolls. The bureaucrats at the Secretariat of the Constituent Assembly were engaged in this activity against the backdrop of the Partition and when more than 18 million refugees were present in the country. This process of turning all adult Indians into voters is however not a legacy of colonial rule. This, as Prof Shani mentions, is a singularly ingenious and indigenous set of decisions undertaken with immense care and an expansive imagination. The scale of this effort included enrolling more than 173 million people on the first electoral rolls. By the time the Constitution of India came into force in January 1950, the abstract notion of universal franchise and the principles and practices of electoral democracy (procedural equality for the purpose of authorising government) were already in place. In effect, Indians became voters before they were citizens.
The concept of instituting universal adult franchise required creating very practical methods which used uniform qualification criteria by which adults will be placed on the rolls. And while for a moment the topic of using the Census activity to create such a roll was discussed, it was decided against as it was clearly understood that these are two different processes and thus would not lead to the outcome desired. Electoral rolls would require creating ways to count the population that makes a break from the established colonial practices of enumeration. The notion that the entire adult population would be considered equal in terms of exercising their voting rights also meant the creation of a framework which set up an all India administrative exercise including the princely states. The model adopted at the time in the State of Travancore provided insights into how to design instructions for the preparation of the rolls. This model included house-to-house visits and registration and thus this continues to be a model that is in place today.
The preparation of the rolls through this process also shows another key aspect of participative democracy - the ability to raise grievances and concerns and seek redress or reforms. Especially in the context of the registration of partition refugees concerns were raised from various places in India to the Constituent Assembly Secretariat. These ranged from exclusionary practices being adopted at the ground level to the more practical problem of the residential requirements being a discriminatory factor against refugees. The process also highlighted the gendered issues with women often being unable to make it within deadlines to register themselves as the officials and state machinery often did not approach this favorably.
As the Secretariat began to respond swiftly and justly to the issues as they come up, there also surfaced topics around migrant labor or adults with no permanent address. As each of these issues received responses and solutions - often far-reaching ones which will have consequences in the years to come - the engagement of the people was exhibited in the passion around the right to vote. Indians took active interest in the processes, and started to relate to the listing of a name on the electoral roll not just as a right but as ‘a title deed to democracy’.
The origin of the single and Central Election Commission has an interesting story. The draft Constitution of February 1948 provided for one election commission for the elections to the central legislature and for separate election commissions for each of the states of the union. However, when Dr Ambedkar presented the article to the House on 15 June 1949 he mentioned that fundamental changes have been brought about wherein the election machinery for all elections to Parliament and the legislatures for every State would be vested in a single Central Election Commission. This change was informed by the experience of attempts at disenfranchisement by local officials on the ground during the preparations of the rolls.
The emergence of the electoral rolls have been made possible through the values and trust elicited by the Secretariat - in the willingness to extensively engage, seek guidance and by ensuring that attempts to disenfranchise groups were resisted and countered. By the end of this process procedural equality for the purpose of voting was bureaucratised. The building of trust among the electorate and the administrators was a precondition for the success of electoral democracy, driven by equality, inclusivity, transparency and accountability.
Prof. Tarangini Sriraman spoke about the history of identity and authentication in India, and why the Indian administration has been obsessed with identifying and criminalizing deviants who are unable to prove their identities.
Tarangini’s talk can be summarized by her concluding statement: “Denials spawn more denials, but inclusion does not spawn more inclusion. If one is a doubtful voter, it is more likely that you will be found guilty in the foreigner’s tribunal, but not the other way around.”
Below are some of the points Tarangini discussed in the talk:
- Authentication is a broad subject that includes passwords, fingerprints, RFID tags, bar codes and numerous others.
- Aadhaar is a cloud-based biometric process which has two steps - authentication of the person and of the process. The enrolment agent will be registered with the CIDR (Central Identities Data Repository). Aadhaar has a multifactor authentication system including the photograph, fingerprint and iris scan.
- Even though Aadhaar enrolment centers are meant to be fully secure and non-replicable, fake centers have been found which have sometimes been given a rubber stamp of approval by the persons in-charge of the real centers.
- The seeding of Aadhaar cards to gas cylinder connections was thought of as a way to solve the problem of duplicates, but this has not worked. Yet another strategy is needed for deduplication.
- While the government was striving for purity and purification of the datasets by cross referencing various databases, in the attempt to build a techno-utopia, 30 lakh people found that their names were missing from the electoral rolls in Telangana.
- The pioneers of fingerprinting were influenced by a particular case of identity theft and were concerned with protecting private property and safeguarding against impersonation.
- William Herschel was interested in studying how fingerprints remain ‘truthful over time’. He recommended the fingerprinting of ryots – so they could not run away to other areas and of criminal tribes.
- Photography was thought to be not as accurate as fingerprinting. It was upheld as a supplement to aid in identification, where the main components were written descriptions, notes of tattoos, moles, forehead measurements etc. These were used to confirm the identities of those convicts who were being sent to the Andaman and Nicobar Islands.
- ‘Kipande’ was a moving pass system in Kenya. It allowed black people to enter white-only areas as it proved that they were employed there. It was criticised as it contributed to the surveilling of black labour. There was organised dissent against similar passes in South Africa. Even after it was universalized, i.e. the passes issued to whites as well, biases still remain. Black people are profiled more often.
- Family Ration Cards became the standard in the 1950s. Strengthened ration cards were suggested by a governor of Bombay and then universalized.
- It was thought that vulnerable documents could be made less vulnerable if the security measures (design) were changed constantly.
- Ration Cards played a major role post-partition. Refugees could get housing after presenting their ration cards. However, the same cards were looked at with suspicion if they were held by a single woman or a Dalit. In this way, documentation does not remove existing biases.
- The State Resident Data Hub (SRDH) database was exploited by a political party in Andhra Pradesh to carry out voter profiling.
- The NRC in Assam caused many problems, including archival panic such as people thronging the state archives to look for proof of their grandparents’ residence. A smudge or a difference in the spelling of names in documents complicated matters,
To reiterate identification has been expected to create more inclusion, but in fact ends up with more denials and doubtful states. As a classic example, the first electoral rolls from Bihar and Rajasthan had nearly 500,000 eligible women voters struck off the list because they had provided the names as a relationship/kinship model viz. wife of <so and so>, mother of <so and so>. This is because traditions prevented them from stating their names to a stranger i.e., the volunteers from the Election Commission of India (ECI). It is an interesting consequence of a requirement that an individual needs to be identified and yet kinship-based names make the system and processes vulnerable and prone to exclusion.
- Nanjala spoke about the intersections between information, technology and the election life-cycle, specifically how manipulation of information corrupts a democratic process. There are numerous examples of this in the Kenyan context: the use of ID card information as a tool for ethnic profiling, and registering members of different ethnic groups as members of political parties without consent. This is a blatant breach of privacy. Such information has also been misused to target people during election-related riots. Similarly information, or lack thereof, can disenfranchise voters as the Kenyan voter ID requires father’s name.
- Due to lack of indigenous capacity, a French company was tasked to build the biometric-based voting system. Ironically, the demographic information of Kenyan citizens now belongs to this French company. Demographic data, augmented with census data, can feed into a surveillance system and be used to intimidate critics of the state.
- Social media has emerged as a double-edged sword. While political parties spread misinformation and hire companies for the same, citizens have used social media to make demands for accountability, leading the government to impose a ban on social media.
- The key takeaway, as Nanjala articulated, “The intersection of technology with a society will always replicate or intensify the underlying fractures from before: if it’s exclusion, technology is only going to intensify it. If it’s misogyny, technology will intensify it. Tech is an intensifier. It’s not a panacea that’s going to fix all of the issues.”
His talk on possible improvements to the voting systems brings up three important topics:
- There already exist reasonably simple methods to handle management of elections including maintenance of electoral roles as well as the votes. So, it could be safely stated that not only are blockchain-based systems not required, they will actually make the process insecure. A permissioned blockchain with only a single trustee (in this case the ECI) amounts to reposing complete trust on it for correctness. Correctnesss can then only be proclaimed but not demonstrated.
- We have to accept that while the VVPAT (Voter Verifiable Paper Audit Trail) was designed to assuage the voter with a visual proof of the vote being cast, it provides no clear method to handle grievances or disputes, or even any certainty that the machine has recorded the vote correctly.
- The ordinary voter may not be familiar with the specifications of ‘end-to-end verifiability’, but they certainly do innately understand that the conduct of an election needs to be manifestly correct.
The cost of an incorrect election is unfathomable! This has been evidenced in many elections where there has been a massive deficit of trust in the governance of the elections as well as the governing authority. And yet, there can be simple systems which are designed to make this process reasonably verifiable.
The absence of standards of maintenance of electoral rolls in India means that often people from marginalized communities find that their names have been deleted from the roles or, it is very difficult to be enrolled as a voter.
To ensure that the ECI is held accountable, a system like a public bulletin board can ensure that anyone is able to see the addition/deletion requests made to the electoral rolls. And this is a much more simple approach than blockchains.
A summary of the session with Srinivas Kodali and Srikanth Lakshmanan on the TSEC e-voting experiment held on 30 October 2021.
Moderator - Sunil Bajpai
The mock elections conducted at Khammam, Telangana, had a process that required that;
The mobile number was linked with Aadhar
The device had OTP auto detection and e-KYC
The name on the Aadhaar matched the name on the EPIC card
The ‘liveliness check selfie’ matched with the photo in the EPIC card
The International Mobile Equipment Identity (IMEI) of the device that registered matched the IMEI of the device on voting day
Just like the Voter Verification Paper Audit Trail (VVPAT) slip which is visible for 7 seconds, after voting the person was shown a screen displaying whom they voted for, for 20 seconds. Then the person receives a message, saying that their vote has been recorded. The app disables screenshots, but it is easy to take a picture of the screen with another phone. It is also possible to connect the phone to a laptop and take a screen recording of the whole process.
This method of voting, in an ideal situation would mean no waiting in lines, negate travelling to the polling station, being able to vote during lunch break at work and most importantly, allow those who live out of the home state to vote. It is believed that this would greatly increase voter turnout.
Telangana has been experimenting with voting methods since 2014. In an attempt to deduplicate records (clean), the databases of Aadhar and voter id were cross-referenced. This resulted in a significantin significant number of people getting dropped from the rolls and being unable to vote in the 2018 election.
The date, 7th October, on which the voting experiment was announced seemed odd. In this period, the parliament had multiple holidays and no discussion of the proposed plan could take place. There was barely any time for outside observers to raise their doubts or ask questions about the procedure. The mock elections were held just 13 days after, on the 20th of October.
During the mock election, Kodali was able to gather people in a space and influence their vote. He was able to ensure by looking at their screens that they were voting for whom he chose. By this he was able to demonstrate that this method does not enshrine the secret ballot as the voting can take place in an unsecured location, without polling officers or observers.
The results of the mock election were not made public, no observers, international or domestic were permitted during the counting, and experts in the field were not invited to test the new method. It was, howeverwas however, deemed a success.
As per the available information the votes cast were to be recorded on a blockchain. A blockchain is thought of as a preferable data store because it is decentralised, the record is stored in multiple places with mathematical algorithms and independent observers, and this makes it less susceptible to tampering by malicious actors. In this case, there was only one viewer/owner, the Telangana State Election Commission, so the implementation of a blockchain could not clearly demonstrate any of the value propositions being relevant and important to the functional completeness of the elections.
Candidate declarations and affidavits are publicly visible, though the link is accessible only for a few days. In case a situation arises wherein the EC removes access, Kaarana (The Kaarana community engages in critical analysis of societal scale technical infrastructure) has created an archive of nominations since 2019.
There were 1400+ app downloads, of which only 12,446 were from Google Play Store, this is worrisome because we cannot be sure that the independently downloaded apps were not versions which were tampered. 3800 people registered on the app. The discrepancy of numbers could be because people who were not eligible to vote also downloaded the app, there is no data to find out how many people were unable to register because of technical difficulties.
Of the 3800, only 2200 votes were recorded. Laxman presumes that this drop is because of technical difficulties like OTP delivery failures, device binding issues (some service providers have particular trouble) and facial recognition authentication issues.
Does this method make voter suppression easier?
One could knock out a cell phone tower to prevent a particular village, which is loyal to a particular party from voting. Voters with bad internet, unreliable devices will be affected. Generators can be blown up as well. Foreign cyber attacks are also possible unless the EC builds its own intranet; this would involve laying fibre cables and is highly unlikely.
Do we need to worry about this method as it is only proposed as a substitute for postal ballots?
Such technology is seen as convenient and could be adopted without proper evaluation. Even the uncertainty of the small percentage that is postal ballots is concerning as a 2016 case regarding the counting of postal ballots in Tamil Nadu was stuck in court for 5 years, the candidate was unable to take his seat until its resolution.
Dr. Taha Ali, during his talk on December 10th 2021, touched upon various discourses and concerns regarding the deployment of electronic and internet voting systems around the world. He has a background in electrical engineering and did further research in network security and electronic voting systems. He is at the forefront of committees engaged in advisory and implementation of election technology in Pakistan. During this talk, he presents a cautionary tale of how discussions and practice of election technology has impacted a developing democracy.
Pakistan has witnessed several socio-economic and political transitions since its independence in 1947. And as a functioning democracy, the 2013 election marks a paradigm shift in the way the country responds to a rigged election, public outrage and the demand for dispute resolution. According to Dr. Ali, the 2013 poll has been historic for three reasons; they are –
For the first time in its history, Pakistan had a transition from a democratically elected government to another democratically elected government. Until then, democratic Governments were overthrown by a military coup, or they were dissolved midway for new elections. In another first, Tehreek-e-Insaf led by Imran Khan emerged as the third front in a country which had always seen a two party competition. The new front also brought an excitement within the voters, proved by the historical turnout of 55% during the polls. The 2013 election was also the most rigged and controversial one in more than 50 years. According to the election watchdog Free and Fair Election Network (FAFEN), more than 71,000 irregularities were reported. From ballot tampering to breach of procedures, there were allegations of massive organized rigging. 21 political parties complained, which led to several threats and abuses to top officials, making them leave their responsibilities in fear.
This unprecedented circumstance led to a severe confusion and distrust amongst the public and the elected government. Dr. Ali then briefed the appalling events that took place in the country following this controversial election.
Later, a judicial committee was formed to investigate the election fraud. This committee concluded that the election was rigged, but it couldn’t find any evidence. So it blamed the Election Commission for the entire organized rigging. Along with this, a parliamentary committee was also formed which introduced the Election Reform Bill of 2017. This bill vouched for an empowered and independent election commission with greater freedom and a transparent results management system. It also wanted to introduce piloting, electronic voting machines and biometric voting machines. The Election Commission (ECP) reached out to Universities for research and modernization of Pakistan’s voting system. Though this discussions continued for more than a year, it didn’t result in any productive action.
Attention then shifted to the Overseas citizens and the measures ECP should take in order to enable Internet voting for them. So in 10 weeks NADRA created a system called IVote which was introduced to the public in 2018. And from the get go, it was heavily criticized because it had repeated thesame problem that bugged the original Australian system which it was modeled after. The problem was so fundamental that it was overlooked so easily; privacy. When you cast your vote remotely, the Election Commission can neither ensure privacy nor voter coercion and identity theft. Malicious websites could fiddle in the process and taking them down from foreign servers was impossible for the Gov.t. The technology on which IVote was built was obsolete and always under threat. Though it was touted to be the world’s largest internet voting system catering to more than 6 million citizens, it failed miserably on arrival.
In response to this, Dr. Ali’s committee recommended three guidelines to go forward with election technology –
Increase investment in R & D in related fields with the help of national and international universities. Great care and deliberation is required to create an ideal structure in which newer methods can function properly. Explore new technologies, look for better and proven methods from around the world. Test the best practices and inform the national discourse on the right direction. They argued that technology was just a component, effort is inbuilding the ecosystem
There are many examples in the developing countries where expensive technology is brought in, but unfortunately doesn’t work in their context. Kenya for instance, introduced a biometric voting machine and deployed them all around the country. On Election Day, the machines had no battery left in them and there were no power sockets to plug it in. Such mistakes are the result of the ‘rush into action’ attitude before proper research and testing.
The committee Dr. Ali is part of was on the forefront of such discussions with the Pakistan Gov.t, but eventually they were sidelined when it became clear that this required a colossal amount of effort. They have teamed up with students and volunteers and raised research funding independently. Though NADRA has recently recognized end-to-end verifiable voting, there is a huge gap in awareness and Pakistan still lacks an authoritative organization which can clearly talk about election technology.
More recently, Pakistan Gov.t has amended the Election Act of 2017 and mandated the deployment of electronic voting machines and internet voting in the next general elections of 2023. The Gov.t has bulldozed it through the resistance from opposition and civil rights groups. It is rushing into the largest deployment of internet voting in the world. With just one and a half year left, it will be impossible to have any feasibility studies, pilot reports, manuals, design specifications or environmental impact studies. The Election Commission or the opposition might challenge this in court, but the decision could go either way. And it might be a disaster on Election Day.
Anonymity is one of the most fundamental aspects in an election and if it is stripped away, so is freedom. A paper ballot inside a box essentially makes it private and secure. Even if someone tries to tamper the box, it would not cost an election, but with internet voting a few clicks may result in thousands of votes going obsolete. An election must be fair and not only that, it should also seem fair. If it doesn’t, then it is a failed election. Like in the U.S., Election systems should be designated as critical infrastructure and an attack on it should be considered an act of war.
In conclusion, Dr. Ali says that “there have been many failed examples of election technology and many that have succeeded. Unsurprisingly, the greatest gains from digitization come from the countries where the quality of democracy is higher and the election commission is independent.”
We should inculcate a democratic culture where technology and its future can be discussed, debated and becomes transparent for the stakeholder.