AI and fake media in Bangladesh’s election ^{CEFR B2}

The February 12 national election in Bangladesh came after a July 2024 student-mass uprising. During the campaign, artificial intelligence became a central tool for political manipulation, shaping what voters saw and believed online.

On December 14, 2025 a photograph circulated that appeared to show Shadik Kayem, the 27th Vice‑President of Dhaka University Central Students’ Union, sitting with a man later alleged to be Osman Hadi’s shooter. Osman Hadi had been shot on December 11. FactWatch confirmed the image was fake and generated by AI.

A study published during the campaign counted 72 cases in which AI-manipulated content gained momentum online. Nearly half of those manipulations created entirely fictional activities or statements (49%), and another 28% used AI to attribute false statements to political figures via edited photocards. Fact‑checkers also found twenty separate cases of AI‑edited fake quotes.

Synthetic images and videos targeted many leaders: AI-made photographs staged campaign scenes for Tarique Rahman, showed Khaleda Zia walking despite reported health problems, and deepfakes depicted former prime minister Sheikh Hasina speaking from exile. An AI video also falsely showed former Malaysian prime minister Mahathir Mohamad endorsing Hasina. Edited photocards mimicked trusted outlets such as Somoy TV and Channel i to increase credibility. The study recorded attacks across parties—forty-seven cases against the Bangladesh Nationalist Party, thirteen against Jamaat‑e‑Islami and six against the Awami League—and described false survey results, staged meetings and persistent claims about Jamaat’s role in Osman Hadi’s shooting. Observers warn these timed escalations, synthetic images and edited photocards are new tactics across the region, and it remains unclear how they will be regulated or how prepared other democracies are to face them.