Bengal SIR issues and truthfulness of the claims of chopping off the voters from the list.

I asked a Grok a question on whether systemic deletion of votes was executed in West Bengal. The response does not make ECI involved in the same. However, use of AI algorithms is being blamed. What does it mean? SIR needs to happen as per the due protocols for securing inclusion and when there are issues due to name mismatch and other algorithm generated issues, sufficient time needs to be there for ensuring correction.

I inserted the following in Grok:

Bengal SIR issues and truthfulness of the claims of chopping off the voters from the list.

Here is the response:

Special Intensive Revision (SIR) of electoral rolls in West Bengal (2025–2026) involved large-scale deletions of voter names ahead of the April 2026 Assembly elections. Claims of “chopping off” voters are factually grounded in the scale of removals, but the truthfulness of allegations of deliberate, targeted disenfranchisement (especially of legitimate voters or minorities) is heavily contested and depends on which figures and interpretations one accepts.

Key Facts on Scale and Process

• Pre-SIR electorate (around October 2025): ~7.66 crore.
• Final figure after SIR and adjudication (by early April 2026): ~6.77 crore.
• Total deletions: Approximately 90.8–91 lakh names (~11.9% of the pre-SIR list).
  • Roughly 63 lakh under the ASDD category (Absent, Shifted, Dead, Duplicate) from the draft rolls (December 2025).
  • An additional ~27 lakh excluded after judicial adjudication of ~60 lakh cases flagged for “logical discrepancies” or as unmapped (no clear link to the 2002 SIR rolls).

The process involved house-to-house enumeration by Booth Level Officers, linkage attempts to the 2002 rolls (treated as a baseline), and an AI/algorithm-driven check for “logical discrepancies.” These included spelling variations in names, age gaps outside expected ranges with parents/grandparents, more than a certain number of voters linked to one ancestor, or gender mismatches with names.

Many of the ASDD deletions (dead, shifted, absent, duplicates) are standard in electoral roll cleaning and are widely accepted as legitimate housekeeping. Analyses indicate only a portion (around 2.4 million in one study) were clearly deceased; the rest involved living people marked untraceable, shifted, or duplicate.

Controversies and Issues Raised

• AI/algorithm and “logical discrepancies”: Critics (including opposition parties, civil society, and some reports) argue the criteria were overly rigid and culturally insensitive. Bengali names often have multiple valid transliterations into English (e.g., Chowdhury/Choudhary, or variations in Muslim names), and minor spelling differences across documents triggered flags. This led to mass notices requiring citizens (including those with passports and long-standing records) to prove eligibility.
• Disproportionate impact claims: Deletions were higher in certain districts (Murshidabad, Malda, North 24 Parganas, Uttar Dinajpur—areas with significant Muslim populations). Some analyses and political claims put Muslim share of deletions higher than their population share (~27% per 2011 Census). Women were also reportedly over-represented in some data. The Election Commission has not publicly broken down final deletions by religion or confirmed large numbers of “foreigners.”
• Timing and process concerns: Conducted rapidly ahead of Assembly polls. Supreme Court intervened, directing special tribunals (19 set up). Of appeals decided by mid-to-late 2026, a very high percentage (~93% in samples of decided cases) resulted in restoration of names, suggesting many exclusions were erroneous or reversible. However, the vast majority of the ~38 lakh appeals remained pending long after the election, so millions could not vote.
• Political claims:
  • Opposition (primarily TMC and allies): Alleged “rewriting” of the rolls, “silent invisible rigging,” or targeted disenfranchisement aimed at their support base (especially Muslims and certain regions). They pointed to the scale, concentration of deletions, and AI flags as evidence of bias. Some described it in strong terms involving citizenship anxieties.
  • BJP and supporters: Framed it as necessary purification of inflated rolls (citing past demographic claims of ~1 crore “infiltrators”/illegal entrants from Bangladesh). They argued the exercise removed ineligible names and that opposition exaggerated the issue for political mileage. Early ECI responses to the Supreme Court called mass-deletion allegations “highly exaggerated” and politically motivated.
  • ECI position: The exercise aimed at pure rolls (no eligible left out, no ineligible included), followed constitutional/legal processes, and included safeguards. It has not disclosed specific numbers of identified non-citizens.

Truthfulness Assessment of “Chopping Off” Claims

• Scale is real: ~91 lakh names were removed. This is not fabricated. A substantial share involved living people (not just the deceased), and the “logical discrepancy” category plus slow/limited pre-poll restoration meant many genuine citizens faced temporary or effective disenfranchisement for the 2026 polls.
• Legitimate cleaning vs. overreach: Routine removals of dead/shifted/duplicate voters are standard and justified. However, the volume of “discrepancy”-based exclusions, reliance on rigid AI rules that ignored common naming practices, concentration in certain demographics/areas, high rate of successful appeals among decided cases, and the inability of most appellants to vote before the election give credible weight to claims of flawed implementation and wrongful exclusions of eligible voters.
• Intentional targeting / “chopping” for political gain: This remains contested and unproven in a definitive sense. Patterns (geography, reported community shares) fuel suspicion among critics, while supporters view it as correcting long-standing inflation. No official ECI data has confirmed large-scale removal of non-citizens matching earlier political claims. Independent analyses and court interventions highlight procedural problems more clearly than proven partisan conspiracy.
• Post-election status: Tribunals continue (slowly). Many restored names indicate reversibility where people could appeal successfully, but the backlog is enormous and the election results were already decided under the cleaned rolls.

In short, the deletions happened on a massive scale, a significant portion went beyond simple dead/duplicate removals due to algorithmic and procedural issues, and many affected people (including documented Indian citizens) were unable to vote in 2026. Whether this constitutes deliberate “chopping off” of legitimate voters for political ends is a matter of interpretation, with strong evidence of over-exclusion and process flaws but no conclusive public proof of coordinated fraud by the ECI. The exercise has left a lasting controversy over electoral integrity in the state.