The Rise of AI-Powered Disinformation in Bangladesh: A Threat to Democratic Processes
Bangladesh is grappling with a surge in AI-generated disinformation, posing a significant challenge to the nation’s democratic processes, particularly as the country gears up for national elections. A report by Dismislab reveals a disturbing trend: the use of sophisticated AI tools, like Google’s Veo-3 technology, to create synthetic videos for political campaigns, marking a significant escalation from previously prevalent image-based misinformation. The Chief Election Commissioner has recognized the severity of this issue, highlighting the potential for AI-manipulated content to sway public opinion and undermine the integrity of the electoral process.
This surge in digitally fabricated content coincides with a broader increase in misinformation, with a 17% rise observed in the second quarter of 2025 compared to the first. While politics remains the primary target, accounting for 44% of all identified misinformation, a notable shift has been the increase in disinformation related to global events. This jump from 1% to 19% coincides with heightened international tensions, such as the Kashmir terror attack and the Iran-Israel conflict, both of which resonated strongly within Bangladesh. The interconnectedness of global events and domestic narratives highlights the vulnerability of information ecosystems to external manipulations.
Election-Related Disinformation: A Growing Concern
The political landscape in Bangladesh, marked by the transition to an interim government and ongoing debates about election timing, has provided fertile ground for the spread of misinformation. Discussions surrounding the potential election date have intensified, with the BNP advocating for elections by December 2025, while the Chief Advisor has suggested a timeframe of February or April 2026. This uncertainty has been exploited, with election-related misinformation seeing an eightfold increase in the second quarter of 2025.
Prominent political figures have become targets of fabricated statements and doctored media, including BNP leader Salahuddin Ahmed, whose image has been used alongside fabricated quotes regarding BNP’s election strategy. False narratives have also targeted Chief Advisor Muhammad Yunus, with claims circulating about him announcing the election date prematurely. Even international figures like US President Donald Trump have been falsely implicated, with fabricated claims suggesting his endorsement of specific political parties.
Beyond Elections: Misinformation Targeting Political Leaders and Activists
The wave of disinformation extends beyond election-related narratives, encompassing accusations of corruption and criminal activity targeting political leaders and activists. Members of the newly formed National Citizen Party (NCP) have been falsely accused of misappropriating funds and resources, while BNP leaders have been targeted with manipulated videos falsely linking them to extortion. Female politicians have faced particularly malicious attacks, with edited explicit content circulating online in an attempt to damage their reputations. This targeted harassment highlights the gendered dimension of online disinformation campaigns.
The Role of AI in Amplifying Misinformation
The use of deepfake technology represents a particularly alarming development in the spread of disinformation. AI-generated videos have been used to propagate false claims, including a fabricated video depicting Chief Advisor Muhammad Yunus’s resignation and another falsely showing Donald Trump praising him. The sophistication of these deepfakes makes them difficult to detect, increasing the potential for them to deceive the public. The fact that even government officials have inadvertently shared such content underscores the pervasive nature of this threat and the urgent need for improved media literacy.
Analyzing the Data: Dismislab’s Methodology
Dismislab’s findings are based on a rigorous analysis of fact-check reports published on eight prominent Bangladeshi fact-checking websites. By analyzing 1,361 fact-check reports published between April and June 2025, the organization identified 1,013 unique instances of misinformation. This meticulous methodology provides valuable insights into the evolving landscape of disinformation in Bangladesh, highlighting the scale and scope of the challenge.
The Need for Countermeasures and Media Literacy
The rise of AI-powered disinformation in Bangladesh demands a multi-pronged approach. This includes strengthening fact-checking initiatives, promoting media literacy programs, and developing technological solutions to detect and counter deepfakes. Furthermore, fostering collaboration between government, civil society organizations, and tech companies is crucial to address this evolving threat and safeguard the integrity of Bangladesh’s democratic processes. Failure to act decisively could have profound consequences for the country’s political stability and social cohesion.