The upcoming elections in Assam have unveiled a disturbing and unprecedented development in the world of political campaigning: the emergence of what’s being called the “first industrialized AI disinformation operation in an Indian state election.” This isn’t just about traditional mudslinging or exaggerated claims; it’s a meticulously crafted, technologically advanced assault designed to manipulate public perception and, horrifyingly, target and dehumanize an entire community. Imagine a political battle where the lines between truth and fiction are not just blurred, but actively erased by sophisticated algorithms and AI-generated content. This report, from a Netherlands-based civil society organization called Diaspora in Action for Human Rights and Democracy (DAHRD), paints a chilling picture of an “altered reality” being manufactured, where a specific community is not only demonized but also systematically excluded from the cultural fabric and even the voting process. It’s a stark reminder that as technology advances, so too do the methods of those who seek to exploit it for divisive political gain, leaving us to wonder about the fundamental fairness and integrity of our democratic processes.
The depth and scale of this AI-driven campaign are truly unsettling. The DAHRD report, titled “AI-Weaponised Disinformation, Systematic Exclusion, and Institutional Failure in India’s 2026 Assam Assembly Election,” meticulously details how 273 social media accounts across Facebook, Instagram, and X, collectively reaching over 407 million followers (a number almost equivalent to the population of the European Union), were weaponized to spread targeted disinformation. This wasn’t a random occurrence; it was a highly organized operation. The report highlights an astounding 432 AI-generated posts on Facebook and Instagram racking up over 45 million views. One Instagram account, “politoons,” alone was responsible for a staggering 88% of all AI content views, generating more than 40 million views from just 102 AI posts. To put this into perspective, with a mere 1% engagement rate, a single post from this network could reach four million people – an astonishing one-sixth of Assam’s registered voters. This isn’t just a handful of individuals spreading rumors; it’s an “industrialized operation” employing synthetic images, deepfake videos, and AI-generated communal content at an unprecedented volume, strategically “front-loaded” before the Model Code of Conduct even came into effect. This means a full-fledged, toxic narrative was already firmly established, making it incredibly difficult for voters to discern truth from fiction before official scrutiny could begin.
The human element of this technological assault is particularly distressing, as it directly impacts individuals and families. The report exposes how thirty-one deepfakes were specifically created to target Gaurav Gogoi, the prominent Opposition face and Congress’s chief ministerial candidate. These deepfakes shockingly portrayed him as a Pakistani agent and a sympathizer of Muslims, a narrative designed to incite distrust and fear. What’s even more alarming is that these fabricated videos were not just circulating in obscure corners of the internet; they were distributed through official BJP accounts and even verified handles belonging to a state cabinet minister. But the vile campaign didn’t stop at political figures. Gaurav Gogoi’s British wife, Elizabeth Colburn, a private individual with no political involvement, was also mercilessly targeted with “six AI-fabricated videos” featuring “intimate and communal scenarios.” The report emphasizes the unprecedented nature of this attack, stating that “no comparable deepfake campaign targeting a political spouse has been documented in any previous Indian election.” This demonstrates a chilling disregard for personal boundaries and a willingness to drag innocent individuals into a political cesspool, all in the pursuit of electoral advantage.
Beyond the direct attacks, the report unveils a broader, more insidious strategy: an “Exclusion Architecture” specifically designed to marginalize and systematically target Assam’s Muslim communities. This wasn’t a singular effort but rather a multi-pronged assault operating simultaneously on several fronts. First, there was the dehumanization campaign, fueled by AI-generated content and, troublingly, 18 verified statements made by Chief Minister Himanta Biswa Sarma himself. Second, the report points to a voter-roll purge that removed 2.43 lakh names, alongside a delimitation exercise that drastically reduced Muslim-majority seats from 35 to a mere 20. This raises serious questions about attempts to disenfranchise a significant portion of the electorate. Third, a “68-post eviction campaign” was hailed as good governance, potentially targeting homes and livelihoods within these communities. And finally, in a blatant attempt to erase cultural identity, there was the “systematic removal of the 17th-century Sufi saint Azan Fakir from Assamese cultural identity.” Perhaps most disturbingly, this AI-fueled propaganda wasn’t just online rhetoric; it directly morphed into enacted legislation within a single electoral cycle. The manufactured narrative of “Land Jihad” was codified into law, restricting property rights—a chilling example of how digital disinformation can translate into real-world legal and social oppression.
The failure of established institutions to address this onslaught of disinformation is a critical concern highlighted in the report. Despite a staggering 119 documented breaches of the Model Code of Conduct (MCC), the Election Commission, the very body tasked with ensuring fair and transparent elections, took “no enforcement actions.” There were no takedowns of platforms responsible for spreading the AI-generated content, allowing the toxic narrative to proliferate unchecked. Even the judiciary, typically a guardian of justice, scheduled its hearings on these critical issues after the polling had concluded, effectively rendering any intervention moot. This institutional inaction creates a dangerous precedent, signaling to those who would exploit AI for political ends that there are few, if any, consequences for their actions. Adding to this disturbing picture, Chief Minister Himanta Biswa Sarma openly acknowledged that he intentionally varied the terminology used in his campaign to evade legal responsibility, even as the targeting remained consistent. His public statement on March 12th, “We had not added the word Bangladeshi—it was constitutionally and legally wrong. But we will correct it and post it again,” offers a stark insight into a calculated approach to circumvent legal boundaries while maintaining the desired harmful message.
The DAHRD report ends with a chilling warning: “Assam is the laboratory. The rest of India is the intended market.” This isn’t just an isolated incident; it’s a dangerous experiment whose findings are now being applied elsewhere. The report notes that in neighboring West Bengal, which is also heading to the polls, millions of voters have been suspended from the voter rolls under a “similar mechanism.” This suggests a disturbing pattern of politically motivated purges and manipulations that extend beyond Assam’s borders. The implications of this AI-weaponized disinformation are profound, extending far beyond a single election. It threatens the very foundations of democratic societies by undermining informed decision-making, fueling prejudice, and eroding trust in institutions. As technology continues to evolve, the need for robust ethical frameworks, vigilant oversight, and swift enforcement against its misuse in political campaigns becomes ever more critical. Without immediate and decisive action, the future of fair elections and the integrity of public discourse may very well be held hostage by algorithms and politically motivated disinformation.