Polls in Nepal become a "digital battleground" due to AI misinformation

Slick AI-generated disinformation has flooded election campaigning in Nepal ahead of Thursday’s vote, the first national polls since deadly protests in September 2025 overthrew the government.

The protests were sparked after authorities briefly banned at least 26 social media platforms, including Facebook, Instagram, YouTube and X, in an attempt to regulate online content. The move triggered unrest led largely by tech-savvy youth frustrated with job shortages and alleged corruption by the political elite. At least 77 people were killed in two days of violence, parliament was set on fire, and the government of four-time prime minister KP Sharma Oli collapsed.

Activists used the messaging platform Discord to propose an interim leader, and days later former chief justice Sushila Karki, 73, was appointed to guide the country to elections.

As Nepal heads back to the polls, social media has again become central to campaigning. Experts and fact-checkers warn that manipulated and fake content is spreading rapidly, especially among first-time voters. Deepak Adhikari, editor of the independent NepalCheck team, said low digital literacy makes it easier for false information to take root. “In a country where digital literacy is low, people believe what they see,” he said.

Technology policy researcher Samik Kharel described the pre-election period as a “digital battleground,” warning that Nepal lacks the expertise to effectively monitor the surge of machine-generated content. Around 80% of the country’s internet traffic runs through social media platforms, he said. According to DataReportal, more than 56% of Nepal’s 30 million population is online, including 14.8 million Facebook users, 4.3 million on Instagram and about 2.2 million on TikTok, based on data from the Internet Service Providers’ Association of Nepal.

AI-generated content has already played a prominent role. Supporters of the ousted premier’s Marxist party shared fabricated drone images showing a massive rally of more than 500,000 people. The images were later reposted by senior leaders. However, fact-checkers at TechPana found they were created using OpenAI’s ChatGPT, and police estimated that fewer than 5,000 people had attended the actual event.

In another case, an AI-generated video circulated on TikTok appeared to show Gagan Thapa of the Nepali Congress urging voters to support a rival party. The platform later removed the video.

Ammaarah Nilafdeen of the US-based Center for the Study of Organized Hate said disinformation poses a serious threat to the integrity of the election process. She also pointed to posts circulating in neighbouring India calling for the restoration of Nepal’s deposed Hindu monarchy, amplified by Hindu far-right supporters online. Such ideological campaigns, she said, contrast with domestic demands for stronger democratic institutions.

The Election Commission has reported widespread use of hate speech and deepfake videos, some created with easily accessible AI tools that show candidates insulting opponents or using obscene language. Information officer Suman Ghimire said more than 600 cases have been referred to authorities, with around 150 handled by police.

In one instance, police detained pro-royalist supporter Durga Prasai over social media posts allegedly aimed at intimidating voters. While the Election Commission has the authority to impose fines or bar candidates, experts say the scale of misinformation online far exceeds enforcement capacity.

Basanta Basnet, editor-in-chief of news website Onlinekhabar, which works with Nepal FactCheck, warned that misinformation could undermine democracy itself. “Candidates and people close to political parties not only compete to win, but also compete to spread misinformation,” he said, adding that false narratives risk pushing citizens to make decisions based on manipulated content rather than facts.