Synthetic Politics and Counter-Framing: AI-Generated Misinformation and Fact-Checking in Bangladesh's 2025-2026 Election Cycle

Abstract

This study examines how artificial intelligence (AI)-generated and digitally manipulated political misinformation circulated in Bangladesh’s pre-election information environment between August 2024 and February 2026, and how fact-checking organisations responded through counter-framing. Combining qualitative framing analysis with descriptive content coding, it draws on the published aggregate datasets of nine Bangladesh-focused verification organisations, covering 4,131 unique misinformation items in 2025 and 1,185 political cases during the 2026 election cycle, alongside a register of verified open fact-checks. The findings show that synthetic media did not replace older manipulation techniques. Instead, it combined with recycled footage, fabricated photocards, fake quotations and partisan amplification to form a hybrid misinformation system. Around 10% of unique items in 2025 involved AI, and roughly 12% of election-related cases were AI-generated, most of them video. The individuals most frequently named included Sheikh Hasina, Muhammad Yunus, Sharif Osman Bin Hadi, Tarique Rahman and Hasnat Abdullah, while the Bangladesh Nationalist Party (BNP) and the interim government carried the heaviest party-level burden. A directional reading shows that content around Sheikh Hasina was mostly favourable, whereas content around the interim government was mostly hostile. The paper argues that AI shifted misinformation from evidential falsification toward synthetic plausibility, making counter-framing dependent on forensic verification and rapid public explanation.