There is the possibility of a small rogue code (perhaps a mere ten lines long) and almost undetectable being inserted by one/two persons in the know. This can be done either through modification of the source code, doctoring the compiler or processer/ ROM, or through the program that controls the VVPAT (Voter Verifiable Paper Audit Trail) machine. Therefore, the presence of the rogue is likely to be in every machine and remain undetectable. The functionality of the rogue is (possibly) that if electronic voting machines (EVMs) are queried after some (say 7) days, it gives biased results to favor a particular candidate. The machine (rogue) identifies the favored candidate as the candidate who secures the majority say (7 votes in the first 10 votes cast). Otherwise, when queried before say week after, it gives unbiased (correct) results. Hence, all checks of the Election Commission would not be able to detect the rogue. Neither would tests prior to polling. All that the person in the know needs to do is urge his/her party workers to be the first to vote at every booth. hen that party gets a bias of say 5-10% in its favor. The current processes of the Election Commission of India (ECI) would not be able to stand in the way of the planting of such a rogue. The EVMs are not subject by the ECI with any large-scale delayed counting tests. Paper trails being counted only on a sample basis would not be able to reveal the rogue. EVM design, manufacture, and procurement of components are not parallelized, or duplicated, and distanced from the government of the day in genuinely autonomous bodies. Their production by public sector entities, which have scant autonomy from the government, introduces a conflict of interest as well.

At a more fundamental level, even if the probability of corruption as described above is very low, the consequences due to the possibilities of “wholesale corruption” are terrifying to a society committed to democracy. The impacts on society are also irreversible, even if they are known many years later. Furthermore, there are strong path dependencies in the sequence of regimes. Hence, it becomes imperative to move back to paper voting, with now excellent checks possible due to CCTVs, human presence, and traceability. Above all, it is important to accept the near-zero probability of “wholesale corruption” of the mandate that paper ballots provide and reject EVMs since they have a non-negligible probability of “wholesale corruption”. Of course, even paper ballots can be corrupted through the use of force/ gendarmes, etc., but then we are talking of “banana republics” where in the first place, there was no intent to be democratic – only to indulge in “mimicry”. Even then, paper ballots can make it known that the voting process has been violated.