EC refuses to give details of SIR software, cites ‘trade secrets, IPR'

The Election Commission declined to disclose the software, application, algorithm, or AI platform involved in identifying errors and anomalies in electoral rolls during SIR. It treated the requested information as protected by commercial confidence, trade secrets, and intellectual property under the RTI Act. This matters because the tools used to flag records can affect which voters receive scrutiny.

M A Mujeeb Ayyub asked whether automated systems flagged records for verification, correction, or deletion. He also sought the workflow after a record was flagged and asked about artificial intelligence, machine learning, predictive analytics, and automated decision support. The Commission did not provide those details.

The Commission separately stated that no new software had been developed or procured for SIR. It said Electoral Registration Officers decide inclusion or exclusion using facts and records. Therefore, the public still does not know which existing tools, if any, supported the exercise or how their outputs were reviewed.

A Special Intensive Revision, or SIR, is a thorough exercise to examine electoral rolls and correct them. It can involve checking whether entries are accurate, identifying inconsistencies, and updating records when the evidence supports a change. This matters because electoral rolls determine which people are listed as eligible voters. The article names SIR but does not provide a complete procedural definition.

In this case, the Election Commission said SIR involved identifying errors, anomalies, or inconsistencies in electoral rolls. M A Mujeeb Ayyub asked whether software automatically flagged records for verification, correction, or deletion. The Commission replied that Electoral Registration Officers decide inclusion or exclusion based on facts and records.

The article does not state the number of voters reviewed, the precise documents checked, or the complete workflow. It establishes that SIR included scrutiny of electoral-roll entries, but leaves the operational details unclear. Understanding those details would help voters know how records are examined and challenged.

The statement means the Election Commission says it did not create or buy a new software system specifically for SIR. That narrows the claim about the exercise, but it does not identify the tools actually used. The distinction matters because an existing government system, application, or analytical process could still have supported checks of electoral-roll data.

Ayyub asked whether software or an AI platform automatically flagged records for verification, correction, or deletion. He also asked about machine learning, predictive analytics, and automated decision support. The Commission did not answer those questions directly. Instead, it said the requests sought clarification outside the information covered by Section 2(f) of the RTI Act.

The remaining uncertainty concerns the system’s identity, capabilities, and workflow. The article does not establish that AI or automated deletion was used. It shows only that no new software was developed or procured, while the Commission maintained that EROs make the final decisions based on facts and records.

The Election Commission said Electoral Registration Officers, or EROs, decide whether an elector is included or excluded from the electoral roll. They are supposed to base those decisions on facts and records. This places formal responsibility with officials rather than with an automated system, a distinction that matters when records are challenged.

Software could still support the process before that decision. For example, it might compare entries, detect inconsistencies, or flag records for verification. Ayyub specifically asked whether records were automatically flagged for verification, correction, or deletion and what happened afterward. The Commission did not disclose the relevant software or workflow.

The article therefore confirms human decision-making but not the extent of technological assistance. It also does not show whether any automated recommendation was accepted, rejected, or independently checked. Future transparency would depend on knowing what systems generated flags and how EROs evaluated them before making decisions.

The article does not state how many voters are represented on India’s electoral rolls. Recent official election figures put India’s electorate at roughly one billion people, so that is the appropriate broad scale rather than an exact figure from this article. The size matters because a small error percentage can become a very large headcount.

If one percent of a roll containing about one billion voters were incorrectly flagged, the number would be about ten million people. A smaller rate, such as one-tenth of one percent, would still represent about one million records. These are illustrations, not figures reported for the SIR exercise. The article provides no measured error rate.

The Commission said EROs make inclusion and exclusion decisions using facts and records. That human review may reduce mistakes, but the article does not describe its safeguards or accuracy. Because the rolls are so large, identifying the tools, flagging rules, and review process would help assess whether errors remain limited or widespread.

If an automated system incorrectly flags a voter’s record, that person could be sent for verification, have information corrected unnecessarily, or face deletion from the electoral roll. Any of these outcomes could create delays, confusion, or a risk of losing access to voting unless the record is restored. The article does not report that such harm occurred.

Ayyub asked whether software automatically identified records for verification, correction, or deletion. He also asked for the workflow followed after a record was flagged. The Election Commission responded that EROs decide inclusion or exclusion based on facts and records. This indicates that a human decision is supposed to follow any possible technological screening.

The unanswered questions concern accuracy, notice, evidence, and review. The article does not describe safeguards, appeal procedures, or the number of records affected. Greater disclosure about the tool and workflow could help show whether incorrect flags are caught before they change a voter’s status.

An algorithm is a set of instructions for processing information or producing an outcome. Source code is the human-readable programming text that implements software. Intellectual property is a broad legal category covering creations such as software, inventions, and protected designs. A trade secret is valuable confidential information kept secret by its owner. These concepts are related but not identical.

A government or contractor might restrict access because disclosure could reveal proprietary technology, weaken security, or provide competitors with protected information. In this case, the Commission cited commercial confidence, trade secrets, and intellectual property when refusing to identify the software and algorithm. EVM manufacturers had also refused earlier RTI requests for source code.

Transparency laws often balance public access against protected commercial or security interests. The article says the Commission treated the requested details as exempt under the RTI Act. It also says some questions were merely clarifications outside Section 2(f). That leaves the public without the tool’s name, version, and detailed workflow.