South Korea’s law enforcement agencies are taking proactive measures against the rising threat of deepfake videos by developing a new detection tool. The Korean National Police Agency (KNPA) plans to deploy this software, specifically designed to identify manipulated video clips or image files, during criminal investigations. Unlike existing detection tools, which primarily rely on Western-based data, this new model was trained on a vast dataset comprising 5.2 million pieces of data from Korean sources, ensuring its effectiveness in detecting deepfakes tailored to Korean contexts.
The deepfake detection tool, developed by the National Office of Investigation (NOI), boasts an impressive 80% probability of accurately detecting whether a video has been artificially generated using AI technology. Equipped with state-of-the-art AI models, this software can analyze video content within five to ten minutes, providing law enforcement with timely insights during investigations. While the tool will play a crucial role in verifying the authenticity of multimedia evidence, authorities emphasize that it will complement traditional investigative techniques rather than serve as standalone evidence.
The introduction of this new technology comes at a critical juncture as South Korea prepares for legislative elections scheduled for April 10. With the campaign period marred by a surge in AI-powered misinformation and deepfake videos, the country’s National Election Commission has reported numerous instances of AI-generated media content violating election laws. These deepfakes, predominantly circulated on social media platforms, aim to manipulate public opinion by distorting speeches or fabricating statements from political candidates, posing a significant challenge to electoral integrity.
In response to the heightened threat of deepfake manipulation, South Korean authorities are not only leveraging advanced detection tools but also collaborating with AI experts from academia and industry. By cross-checking detection results and consulting with specialists, law enforcement seeks to minimize the risk of erroneous identifications and ensure the accuracy of investigative outcomes. Ultimately, the deployment of this deepfake detection tool represents a proactive effort by South Korea to safeguard its democratic processes and combat the proliferation of misleading information in the digital age.
