SEOUL — As South Korea approaches its parliamentary elections in April, the rise of artificial intelligence (AI)-generated deepfake videos and images is causing concern among election authorities. With less than 50 days until the elections, the National Election Commission (NEC) is on high alert, detecting a significant increase in AI-generated content that violates the country's electoral laws.
According to Yonhap News Agency, since January 29, when the revised election law came into effect, 129 instances of AI-generated media content have been identified as illegal. The updated law, enacted to address the growing threat of deepfakes, imposes penalties of up to seven years in prison or fines of up to 10 million won (approximately US$7,500) for the use of deepfakes in election campaigning.
The NEC's efforts to combat deepfakes are part of a broader initiative to protect the integrity of the electoral process from the risks posed by advances in AI technology. The revised legislation, passed by the National Assembly in December, seeks to curb the spread of misinformation and uphold democratic values amid emerging digital threats.
Deepfakes have been employed in previous elections, both in South Korea and globally, to create fake endorsements and spread false information. For example, an AI-generated video that falsely showed President Yoon Suk Yeol endorsing a local candidate was circulated during the 2022 provincial elections. Similarly, voters in New Hampshire, USA, received an AI-generated robocall in January, impersonating U.S. President Joe Biden and discouraging participation in the presidential primary election.
The rapid development and sophistication of deepfake technology present significant challenges for election authorities. The production of realistic AI-generated content can outpace the verification processes currently in place, making it difficult to promptly identify and mitigate these threats.
Kim Myuhng-joo, a professor at Seoul Women's University, highlighted the disparity between the creation of deepfakes and the NEC's ability to verify them, stating that the speed at which believable fakes can be produced far exceeds the commission's verification capabilities. The NEC acknowledges the evolving nature of deepfake technology and emphasizes the importance of a comprehensive verification process to address potential legal violations effectively.
As the elections draw nearer, the NEC and other stakeholders are working to refine their strategies for detecting and responding to deepfake content, underscoring the need for vigilance and innovation in safeguarding democratic processes.