ADR v. Election Commission (2024): Supreme Court upholds EVM-VVPAT, refuses 100% verification but tightens safeguards

Few institutional questions in Indian public life have proved as durable as the trustworthiness of the electronic voting machine. By 2024, EVMs had been used across the country for more than two decades, paired since 2013 with the Voter Verifiable Paper Audit Trail (VVPAT) — a printer that displays, then drops into a sealed box, a paper slip recording each vote. Petitioners led by the Association for Democratic Reforms (ADR) asked the Supreme Court to go further: to require that every VVPAT slip be counted against the EVM tally, or, failing that, to return to paper ballots altogether. In Association for Democratic Reforms v. Election Commission of India (2024 INSC 341), a Bench of Justices Sanjiv Khanna and Dipankar Datta declined both prayers, but used the occasion to harden the procedural guarantees around the machines.

The facts in brief

The petition, Writ Petition (Civil) No. 434 of 2023, was heard on the eve of the 2024 general election. At that time, the Election Commission's protocol required that VVPAT slips be hand-counted and matched against the EVM electronic count in five randomly selected polling stations per assembly constituency or segment — a sampling regime put in place after the Court's earlier intervention in N. Chandrababu Naidu v. Union of India (2019), which had raised the figure from one booth to five.

ADR and the other petitioners contended that five-booth sampling was statistically inadequate to detect manipulation and that voters had no way of confirming that the slip they saw printed actually corresponded to the vote recorded electronically. They sought either count-every-slip verification or a reversion to ballot papers. The Election Commission and the EVM manufacturers explained the architecture of the machines — the control unit, the ballot unit, and the VVPAT, each with a one-time-programmable burnt memory microcontroller — and pointed to the absence of any demonstrated instance of EVM tampering across millions of machines and audited slips.

The question

Two questions framed the dispute. First, does the constitutional guarantee of free and fair elections under Article 324, read with the right of a voter to know that their vote is correctly recorded and counted, require 100% verification of VVPAT slips against the EVM count? Second, if not, are the existing safeguards sufficient, or must the Court calibrate additional procedural protections to sustain public confidence in the machines?

What the Court held

The Court answered the first question in the negative. It held that the EVM-VVPAT system was secure and that the petitioners had produced no evidence of actual tampering; suspicion and theoretical vulnerability, untethered from proof, could not justify dismantling a system that had functioned reliably. Justice Khanna's opinion stressed the practical and administrative consequences of total verification — delay, cost, and the risk of human error in mass manual counting — against speculative gains in accuracy.

Having upheld the system, the Court turned to the second question and issued two concrete safeguard directions. First, the Symbol Loading Units (SLUs) — the devices used to load each candidate's ballot symbol into the VVPAT — must, on completion of the symbol-loading process undertaken on or after 1 May 2024, be sealed and secured in a container, with candidates or their representatives permitted to sign the seal, and stored in the strong room alongside the EVMs for at least 45 days after the declaration of results. This closes a gap the petitioners had highlighted, since the SLU is the point at which symbol data enters the machine.

Second, the Court created a post-result audit on demand. Candidates placed second or third in a constituency may, within seven days of the declaration of results, make a written request for verification of the burnt memory or microcontroller of 5% of the EVMs (the control unit, ballot unit, and VVPAT) per assembly constituency or segment. The check is to be conducted by the manufacturer's engineers, in the presence of the candidates or their representatives, at the requesting candidate's expense — but the cost is to be refunded if any tampering is found. Justice Dipankar Datta wrote a separate, concurring judgment agreeing with the result and the directions.

Analysis

The judgment is best read as an exercise in institutional deference disciplined by remedy. The Court refused to second-guess the Election Commission's technical choices or to substitute its own view of the optimal verification rate — a stance consistent with its general reluctance to micromanage the machinery of elections. Yet it did not treat deference as the end of the inquiry. By converting two of the petitioners' anxieties — the integrity of symbol loading and the unavailability of any post-result forensic check — into enforceable directions, the Court acknowledged that confidence in elections is itself a constitutional value that can be served by procedure even where substantive relief is denied.

The 5% microcontroller-check direction is the more novel of the two. It gives losing candidates, rather than the Commission alone, a trigger for forensic examination of the silicon at the heart of the machine, and the refundable-cost design is a deliberate filter: it deters frivolous demands while ensuring that a candidate with a genuine grievance is not out of pocket if tampering is proven. It is, in effect, a fee-shifting mechanism imported into electoral audit.

The Court's reasoning also marks a continuity with its earlier electoral-machinery jurisprudence. Where N. Chandrababu Naidu had expanded sampling from one to five booths as a confidence measure, ADR consolidates that approach: incremental, evidence-sensitive safeguards rather than wholesale redesign. The repeated insistence that doubt "without supporting evidence" cannot drive constitutional remedy is the doctrinal spine of the decision.

Why it matters

ADR v. ECI is the Supreme Court's most authoritative recent pronouncement on the constitutional adequacy of electronic voting, and it will be cited whenever the EVM debate resurfaces. For election lawyers, the operative takeaways are the two directions: the 45-day SLU sealing now forms part of the audit trail available to candidates, and the 5%-microcontroller verification route is a new, time-bound post-result remedy that must be invoked within seven days. The judgment does not foreclose future challenge — it expressly rests on the absence of proof of tampering — but it sets a high evidentiary threshold for anyone seeking to displace the EVM-VVPAT system, and it signals that the Court will respond to credible concerns with calibrated procedure rather than abolition.

Related on Valkya

• ADR v. Union of India: Electoral Bonds
• ADR v. ECI: Special Intensive Revision of Electoral Rolls
• PUCL v. Union of India: Voter's Right to Information
• Mohinder Singh Gill v. Chief Election Commissioner

Sources

• Supreme Court Observer, "VVPATs for Voter Verification — Association for Democratic Reforms v. Election Commission of India"
• Verdictum, "EVMs Are Simple, Secure & User Friendly; Repeated Doubts Without Evidence Can Create Distrust: SC Rejects Plea For 100% EVM-VVPAT Verification"