Let process, not tech, settle EVM debate

India's Electronic Voting Machines (EVMs) are again at the centre of a controversy following recent electoral outcomes. Opposition parties, particularly after facing defeats, have raised serious concerns about these machines' potential vulnerability to manipulation. The primary concerns about EVM manipulation typically focus on sophisticated hacking or technological interference. These are legitimate technical questions that both the Election Commission of India (ECI) and manufacturers have addressed extensively. However, drawing from my experience as a presiding officer of a polling booth in the 2023 Karnataka Assembly election, it is evident that the integrity of Indian elections depends not only on the infallibility of machines but also on the meticulous execution of established protocols and the vigilance of electoral stakeholders.

The electoral process incorporates multiple layers of security and verification, beginning well before the first vote is cast. Some have raised the possibility that votes could be programmed to register differently than intended. The system addresses this serious technical concern through multiple verification steps. Each polling day starts with a mock poll, at approximately 5.30 am, where polling agents of the candidates can systematically test the machines by pressing each candidate's button in sequence. This process generates verifiable VVPAT (Voter Verifiable Paper Audit Trail) slips that agents can inspect before securing them. Only after the results are tabulated and accuracy confirmed does the actual voting begin. In the Karnataka polls, polling agents tested every candidate's button functionality, creating a verified record of proper operation and zero votes before the start of voting.

The voting process itself is structured for constant verification. There is a periodic record and release of turnout throughout the day. Polling agents possess the right to maintain their own electoral roll records, tracking each vote cast. Voter identifications are announced publicly, enabling real-time verification. The VVPAT system provides a seven-second window for each voter to verify their selection, accompanied by an audible confirmation beep. This ensures that voters know that their vote was registered and can immediately challenge any discrepancy. There were no such challenges in my booth.

Some critics suggest votes might be added after the polling ends. However, once polling is declared closed and the “close” button is pressed on the EVM, the machine physically cannot accept more votes without being reset. Moreover, the control panel is sealed after morning mock polls, with each seal requiring polling agents' signatures. The machine then undergoes three distinct levels of sealing: the interface panel, the machine itself, and finally, the carrying case. During the Karnataka elections, polling agents verified and signed each seal, making it impossible to substitute machines during counting without detection.

Much has been made of EVMs showing high battery levels at the day's end, particularly in booths favouring certain parties. The ECI's explanation is straightforward – machines designed for 2,000 voters and four ballot units naturally show minimal drain when handling fewer voters with single units. More importantly, the battery level debate misses a crucial point: agents can verify machine authenticity through their signatures on numbered seals, making battery levels irrelevant to the question of tampering.

The possibility of remote manipulation or post-poll tampering is also addressed through a system of multiple procedural checks. First, Form 17C, provided to all polling agents at the end of the polls, serves as a crucial independent record. This document captures the total votes cast, unique identification numbers of the EVM and the identification number of the seals used. When votes are counted at the counting centre, these numbers must match with the EVM count. Second, throughout the polling day, agents maintain their own real-time count that must reconcile with both the Form 17C and the final EVM count. Third, the VVPAT system provides physical proof of each vote through paper slips.

Protocol non-compliance

Despite these safeguards, ground-level observations reveal a concerning pattern: polling agent participation in these crucial procedures often falls short of expectations. In the case of the Karnataka polls, few candidate agents were available throughout the poll, and none of them had any training on what they had to do. For example, during the mock poll, each candidate's button must be tested, with agents verifying the VVPAT slip and electronic counts. The process is thorough, but its effectiveness depends entirely on agent participation and understanding. I had to explain to the present agents that they needed to verify each slip, count the totals, and sign the verification forms. None of this was intuitive to them; they were simply going through motions they didn't fully understand.

The sealing process further illustrates this gap between procedural robustness and human execution. Each EVM undergoes multiple sealing steps, with each seal requiring agents' signatures. As a Presiding Officer, I had to repeatedly remind agents to verify the seals before signing. Most were ready to sign without checking, missing the point that their signatures were their candidates' guarantee against machine tampering. They seemed unaware that at counting, they or their colleagues would need to verify these very signatures.

Perhaps the most telling was the handling of Form 17C – the crucial document recording total votes cast, voting times, and machine numbers. When polls closed, most agents were ready to leave immediately. I had to insist they remain to verify the numbers and collect their copies of Form 17C. None seemed to understand that this document was their primary safeguard against post-poll manipulation.

While these observations come from my experience in a single polling booth in Karnataka, and practices may vary across locations and elections, they illustrate that the real vulnerability in India's electoral system may lie not in its technical elements but in the lack of effective human oversight. Political parties would better serve democratic interests by investing in comprehensive training for polling agents, ensuring they understand their rights, responsibilities, and their crucial role in maintaining electoral integrity. Strengthening these human elements represents the most effective path to ensuring the continued integrity of India's electoral system.

(The writer is an Assistant Professor of Law at NLSIU, Bengaluru)