National Forensic Service
AI analysis of suspected deepfake evidence
South Korea's National Forensic Service deployed AI deepfake analysis and reported examining 53 election-related videos
Problem
Manipulated images and recordings were difficult to authenticate with predominantly manual forensic methods. Investigators needed help examining suspected deepfake evidence.
Approach
The National Forensic Service developed an AI model to automate analysis of images, video and audio. It incorporated results from government-supported research and applied the technology to investigative work.
Results
Takeaway
The service reported election-related forensic examinations, but did not disclose detection accuracy or a measured reduction in examination time.
Categories
- Industry
- Public sector
- Business function
- Production and quality
- Innovation type
- Process innovation
- AI technology
- Computer visionPredictive analytics and ML
- Adoption
- Built in-house
- Evidence
- Company's own claim
Sources
Source summaries
Dailian reported the agency's internally developed AI deepfake analysis model and its application in investigations. It cited the agency's experience examining 13 cases involving 53 videos during the presidential election.
Full details are in the original sources.