Deepfake Interest in South Korea: A Temporal Analysis of Google Trends from 2017 to 2024

Deepfake technology, which utilizes artificial intelligence to generate hyper-realistically manipulated videos, images, texts, and audio, has garnered significant public and academic interest. The proliferation of deepfakes, especially in non-consensual pornography, financial fraud and political misinformation, has sparked ethical, moral, legal, and security debates worldwide. While existing research predominantly focuses on deepfake detection, legal frameworks, and their potential impact on the democratic process, few studies have examined public interest in deepfakes and the factors influencing search behavior. This study addresses this gap by analyzing public interest in deepfakes in South Korea, using Google Trends data from January 2017 to August 2024. This timeframe is particularly significant as it encompasses the initial emergence of deepfake technology in 2017 and its increasing use in fraudulent and non-consensual content in South Korea. The country represents a unique case due to its global leadership in deepfake-related searches, widespread consumption of non-consensual sexual deepfakes, and frequent occurrence of deepfake fraud. This study employs dictionary-based text analysis to categorize search queries into three main themes: sexual content, techniques for creating deepfakes, and methods for accessing deepfake materials. The findings indicate that 77.81% of searches are related to non-consensual sexual content, primarily targeting female celebrities. Contrary to global trends, political deepfakes did not significantly influence search patterns in South Korea. These insights highlight the urgent need for stronger regulatory frameworks and technological interventions to mitigate the harms associated with deepfakes.