The chief election commissioner (CEC) has highlighted the significant challenge presented by artificial intelligence (AI) in the upcoming election, labeling it as more perilous than conventional weapons due to its potential to disseminate false information. AMM Nasir Uddin expressed concerns about the ability to manipulate content using his likeness and voice, branding the misuse of AI a contemporary threat capable of interfering with election campaigns and influencing the voting process. He also touched upon other election-related issues like the presence of illegal arms, instilling trust in the electoral system, and boosting voter turnout, earmarking these for future discussions.
While acknowledging the issue is a crucial first step, concrete actions need to be taken. It is imperative for the Election Commission to implement effective protective measures, including setting up robust monitoring systems to detect AI-generated content, partnering with experts and media outlets to mitigate its impact, updating legal frameworks, and enhancing public awareness. The existing electoral code of conduct lacks specific guidelines on this matter, necessitating immediate attention. Although traditional measures to ensure fair competition remain essential, tackling AI misuse takes precedence. This urgency is emphasized by both global and local incidents, with a recent report from The New York Times indicating that AI was employed in over 80% of elections in 2024, influencing at least nine major elections this year.
The ease and credibility of fabrications have escalated with AI, exemplified by a recent study by Dismislab that identified 70 AI-generated political campaign videos produced between June 18 and 28. These videos, forged using Google’s Veo text-to-video AI model, featured entirely fictitious characters such as rickshaw drivers, garment workers, teachers, Hindu and Muslim women, and youth, endorsing various political factions.
Recalling past instances of fake content dissemination, including cloned candidate voices during recent elections, the current proliferation of such deceptive practices signals a heightened threat. The proliferation of emotionally charged synthetic content raises alarms about its potential disruptive impact as the election approaches.
It is evident that stronger measures are required to counter this trend. While eliminating the risk of AI-generated content entirely is unfeasible, and not all such content is malicious, the EC must collaborate with pertinent state entities, political groups, and social media platforms to combat its misuse. Drawing insights from nations where AI has already interfered with elections is crucial. Without prompt and informed actions, Bangladesh risks facing election disruption from such technologies.