India’s 2026 Assam Election Threatened by AI Disinformation Amid Institutional Concerns

The Indian state of Assam has become ground zero for what watchdog groups call an unprecedented campaign of AI-powered disinformation targeting its Bengali-speaking Muslim population, according to a newly released monitoring report.

The Diaspora in Action for Human Rights and Democracy (DAHRD) has published findings documenting what it describes as the most sophisticated AI-enabled disinformation operation ever recorded in an Indian election. Their investigation reveals a coordinated effort combining digital propaganda, legislative action, and administrative procedures ahead of the April 9, 2026, Assam Assembly elections.

According to the report, authorities established a six-tier content ecosystem that produced 432 confirmed AI-generated posts that reached over 45.4 million viewers. Researchers characterized the operation as “industrial-scale” and designed to manufacture an altered reality in which an entire community faces simultaneous dehumanization and disenfranchisement.

One of the most concerning findings highlights what investigators call a “propaganda-to-policy pipeline,” where synthetic content establishing narratives such as “Land Jihad” directly influenced legislation. The report specifically points to a February 2026 property restriction law that prohibits land sales to Muslims in certain designated areas of Assam.

The monitoring group documented an alleged tactical vocabulary shift by Chief Minister Himanta Biswa Sarma, who reportedly rotated terminology from “Miya” to “Bangladeshi” to “Encroachment” while maintaining identical targeting of Muslim communities. According to the report, Sarma acknowledged this linguistic adjustment in a March 12 interview, stating that earlier messaging was “constitutionally and legally wrong” because it had not employed the term “Bangladeshi.”

Beyond rhetoric, DAHRD identified what it calls a systematic “exclusion architecture” operating through four parallel mechanisms: dehumanization via AI-generated content, the deletion of approximately 243,000 names from voter rolls through a Special Intensive Revision process, physical displacement through forced evictions (which were subsequently celebrated on social media), and the cultural erasure of figures like the Sufi saint Azan Fakir from regional historical narratives.

The monitoring report directly implicates high-level political figures in spreading disinformation. Investigators documented 31 confirmed deepfakes targeting opposition candidate Gaurav Gogoi, including a fabricated video shared from a verified Cabinet Minister’s social media account portraying Gogoi as a “Pakistani agent.” The campaign allegedly employed unprecedented gendered disinformation tactics by targeting Gogoi’s wife, Elizabeth Colburn, a private citizen, with at least six AI-fabricated scenarios of an intimate and communal nature.

Despite documenting 119 breaches of the Model Code of Conduct, with 84 classified as high-severity violations, the report notes a complete institutional failure to address the situation. According to DAHRD, the Election Commission of India has taken no enforcement action, while social media platforms have executed zero content takedowns related to the documented violations.

The organization warns that Assam appears to be functioning as a testing ground for these techniques. They note with concern that similar administrative mechanisms have reportedly already been deployed to suspend over 10 million voters in the neighboring state of West Bengal.

In its conclusion, DAHRD cautions that without urgent intervention, the widening gap between AI propaganda production capabilities and democratic accountability mechanisms threatens to further compromise the integrity of India’s 2029 general elections.

The findings raise serious questions about electoral integrity and the evolving role of artificial intelligence in political manipulation, particularly in regions with existing ethnic and religious tensions.