On Thursday, an estimated 128 million Bangladeshis will head to the polls in what is set to be the largest democratic exercise of 2026. For millions of voters, it will be the first genuinely competitive election of their lives.
Coming after the student-led uprising that toppled Sheikh Hasina in August 2024—and nearly two decades of authoritarian rule before that—the stakes could scarcely be higher. A credible vote promises a long-awaited return to democratic legitimacy.
But the election is only half the battle.
A new SecDev report, The Ballot and the Algorithm: Extremism, Elections, and Bangladesh’s Democratic Reset, warns that a parallel contest is already underway—one that will outlast polling day. It is being fought not at ballot boxes, but across digital platforms where Bangladesh’s overwhelmingly young electorate now forms its political views.
The report, produced jointly by SecDev and its local partner Rupantar under a Global Affairs Canada–supported project, paints a stark picture of how online extremism has surged alongside political transition. Between January 2020 and January 2026, subscriptions to violent extremist channels in Bangladesh exploded by 2,071 percent—from 1.7 million to more than 37 million. Over the same period, social media users grew by 77 percent and mobile internet subscriptions by 22 percent.
It says extremism is not growing simply because internet access is expanding, rather existing users are being pulled deeper into radical ecosystems.
Al-Qaeda’s regional affiliate, AQIS, dominates this space, controlling 62 percent of known violent extremist channels. Its objective, the report argues, is not immediate violence but long-term influence—building social legitimacy through religious outreach, exploiting grievances around sovereignty and identity, and steadily eroding faith in democratic pluralism while staying below the threshold of illegality. It is influence built as infrastructure: patient, scalable, and hard to uproot, adds the report.
Since August 2024, it argues, the interim period has proven unusually permissive. Extremist groups have held mass rallies, issued threats against cultural events, and openly positioned themselves for political participation. At the same time, counterterrorism agencies have been sidelined and the terrorism threat publicly downplayed. For these groups, prolonging an interim arrangement is far preferable to facing an elected government with a fresh popular mandate.
As per the report, the most dangerous election risk remains Islamic State–aligned violence: low probability but potentially devastating. The Keraniganj madrasa blast demonstrated that Neo-JMB networks still possess active IED-making capacity, with evidence suggesting preparations for up to 85 devices aimed at election-related targets. Even a single successful—or credibly attempted—attack could disrupt voting, undermine confidence in the results, or justify an extended interim period, particularly in an environment where militant-linked disinformation dulls public awareness of the threat.
If the election proceeds credibly, it will restore a minimum threshold of democratic legitimacy. But it will not dismantle networks that have spent the past eighteen months entrenching themselves online and offline, the report mentions, warning that the post-election phase will mark a shift from disruption to subversion: questioning outcomes, exploiting transitional justice, and reframing every security response as a return to past repression.
Bangladesh is entering a new and fragile phase, it concludes, adding that the question is no longer whether democracy can return—but whether, under pressure from both the street and the screen, it can endure.
