AI and Synthetic Reality in Bangladesh's 2026 Election

• The 2026 Bangladesh general election, the first since the 2024 student-led uprising, was marked by the widespread weaponization of AI and synthetic content to distort political narratives.
• A study identified 72 distinct cases of AI-manipulated content, including deepfakes, synthetic imagery, and fabricated media "photocards" designed to influence voter perception.
• The influx of synthetic reality overwhelmed fact-checking efforts, as emotionally resonant fabrications often overshadowed factual denials.

Tactics of Manipulation

• Synthetic Imagery: AI-generated images depicted political leaders in false scenarios, such as digging fields or shaking hands with children, to manipulate public personas.
• Fabricated Photocards: Political actors created fake graphic news cards mimicking established outlets like Somoy TV and Channel i to attribute inflammatory, false quotes to candidates.
• Deepfake Videos: Sophisticated videos circulated false endorsements—such as a fake endorsement of Sheikh Hasina by Malaysia's Mahathir Mohamad—and purported statements from exiled political figures.

Targeting Political Parties

• Bangladesh Nationalist Party (BNP): Faced the heaviest volume of attacks, with 47 targeted campaigns during their successful landslide bid.
• Jamaat-e-Islami: Subjected to 13 campaigns, including false quotes and misinformation regarding their role in the shooting of activist Osman Hadi.
• National Citizen Party: Youth leaders like Nahid Islam were targeted by doctored media aimed at alienating their core younger voter base.
• Awami League: Despite being banned, the party was targeted via deepfakes of former PM Sheikh Hasina, intended to challenge their remaining legitimacy.

Wider Implications

• The election serves as a blueprint for AI disinformation in South Asian democratic processes, providing a model of how synthetic content can escalate toward polling days.
• Conspiracy theories gained undue credibility through AI-manufactured "evidence," such as fake photos depicting secret meetings between foreign officials and activists.
• The volume and speed of AI-generated content created a post-truth electoral environment where traditional verification mechanisms struggled to keep pace.