Development of AI Forensic Techniques for Tracing the Entire Crime Process

The police have taken proactive measures against evolving cybercrimes such as deepfakes (manipulated videos) by developing artificial intelligence (AI) detection software, leading to the apprehension of 419 individuals over a seven-month period.


The National Investigation Headquarters of the National Police Agency announced on August 13, 2026, that between January and July of this year, 419 suspects involved in deepfake-related crimes were apprehended, with 17 of them taken into custody. In response to increasingly sophisticated crime methods, the police have established and are further advancing a triple-layered response system consisting of: ▲ deepfake detection software identification, ▲ scientific forensic video analyst interpretation, and ▲ analysis of AI tools used in the crimes.


The police have developed an "AI Forensic Technique" that analyzes the unique features of commercial artificial intelligence (AI) programs being misused for deepfakes in the field of digital forensics. A professional analyst is demonstrating the process of analyzing deepfake crimes using AI forensics. National Police Agency

The police have developed an "AI Forensic Technique" that analyzes the unique features of commercial artificial intelligence (AI) programs being misused for deepfakes in the field of digital forensics. A professional analyst is demonstrating the process of analyzing deepfake crimes using AI forensics. National Police Agency

View original image

For example, in March, the National Police Agency's cybercrime reporting system received a series of reports stating, "Sexual deepfake videos featuring my synthesized image have been posted on overseas adult websites." Upon receiving these reports, the agency immediately coordinated with four regional police agencies to secure relevant evidence and requested the removal of offending accounts and posts from the sites in question. Analysis using deepfake detection software revealed common distinctive features. Focused investigation uncovered that the distributed videos were created by the same individual, leading to the arrest and indictment of a man in his twenties.


The software utilized by the police is also in use by several overseas law enforcement agencies. It conducts comprehensive analyses—such as determining whether faces have been generated or swapped, monitoring biological features like eye blinking and blood flow changes, and examining file metadata structures—to identify synthetic content. The system is equipped with AI models specifically trained for detecting deepfakes in Korea and is continuously being enhanced. To date, the software has been deployed in 1,636 cases and is also used in investigations of election-related crimes.


Additionally, the police have completed the development of an "AI forensic technique," which analyzes the peculiarities of commercial AI programs exploited for deepfakes in the digital forensics field. Professional analysts at regional police agencies technically analyze and reconstruct the entire chain of criminal activity—from creation to distribution—by determining which programs the suspect used to manipulate the video and identifying the specific portions altered. These analyses serve as crucial evidence for proving criminal intent.



A representative from the National Investigation Headquarters stated, "We are advancing police response technologies even faster than the rate at which technology evolves," adding, "Through system enhancement and focused crackdowns, we will enforce a zero-tolerance policy in responding forcefully to deepfake crimes."


This content was produced with the assistance of AI translation services.

© The Asia Business Daily. All rights reserved. Unauthorized AI training and use prohibited.

Today’s Briefing