South Korea’s law enforcement agencies are taking proactive measures to combat the rise of deepfake technology in the country. In response to the increasing use of deepfakes in criminal investigations, the Korean National Police Agency (KNPA) has announced the development of a new deepfake detection tool.
The KNPA revealed on March 5, 2024, to the South Korean press agency Yonhap that the National Office of Investigation (NOI) will soon deploy new software specifically designed to identify manipulated video clips or image files created using deepfake techniques. Unlike other AI detection tools that have been trained on Western-based data, this new software has been trained on 5.2 million pieces of data from 5400 Koreans and related individuals. The KNPA emphasized that this model uses the latest AI technology to detect new types of hoax videos that were previously unseen.
The advanced tool is capable of determining whether video content has been artificially generated using AI technology within just five to 10 minutes of analysis. The Police stated that the software boasts an 80% accuracy rate in detecting whether a video is authentic. It will be seamlessly integrated into NOI’s investigative processes, providing crucial insights to aid in criminal investigations. However, the South Korean police have clarified that the tool will guide investigations rather than serve as direct evidence.
To ensure the reliability of the detection process and minimize the chances of false positives, South Korean law enforcement agencies plan to collaborate with AI experts from academia and private companies specializing in electoral disinformation. This collaborative approach will help in verifying the authenticity of deepfake content and strengthen the overall investigative process.
The unveiling of this new deepfake detection tool comes at a critical time for South Korea, as the country prepares for its upcoming legislative elections scheduled for April 10. The election campaign has already been marred by a significant increase in AI-powered misinformation and disinformation campaigns.
In February, the National Election Commission (NEC) of South Korea reported 129 instances of AI-generated media content, including fake videos and audio, that violated newly revised election laws. These deepfakes were primarily circulated on social media platforms, where they targeted opposing candidates by distorting or altering speeches to mislead the public.
The proliferation of deepfake technology poses a significant threat to the integrity of elections and public discourse. By investing in cutting-edge detection tools and adopting proactive measures to counter deepfake threats, South Korea’s law enforcement agencies are demonstrating a commitment to maintaining the trust and transparency essential for a fair democratic process.
As South Korea prepares for the 2024 elections, the implementation of robust deepfake detection measures highlights the country’s dedication to combating online misinformation and safeguarding the electoral process from malicious actors seeking to manipulate public opinion.