DAEJEON, Sept. 10 (Korea Bizwire) — The Electronics and Telecommunications Research Institute (ETRI) said Thursday that it developed a new artificial intelligence (AI) technology to analyze the harmfulness of video content.
After searching websites on the basis of registered keywords, the new AI detects harmful content by analyzing the websites’ text and images.
The AI is 99.4 percent accurate in locating harmful content, and it takes less than 0.01 seconds to detect offensive material.
During a two-month trial that kicked off in June, the new AI technology automatically searched a total of 18,945 websites using content search keywords, thumbnail images and URL.
Among them, a total of 2,631 websites were identified as having harmful content, and successfully filtered.
Lee Nam-kyung, head of ETRI’s medial intelligence research department, noted that the technology could be employed to support victims of digital sexual crimes.
J. S. Shin (firstname.lastname@example.org)