TAIPEI (TVBS News) — The Ministry of National Defense (MND, 國防部) on Tuesday (July 16) invited the National Chung Shan Institute of Science and Technology (NCSIST, 中科院) to present a special report on deepfake recognition, examining the technology and its defense applications.
This marks the first time the ministry has hosted such a discussion, highlighting the growing concern over deepfake technology in national security.
NCSIST's Information and Communications Research Division director Huang Yi-chin (黃益進) emphasized that artificial intelligence (AI) offers innovative uses and convenience, but deepfake technology, which creates realistic fake content, poses significant risks. "Deepfake technology is often used in information warfare to achieve the goals of cognitive warfare. Both deepfake and altered satellite images can create non-existent buildings or terrain," stated Huang.
Huang explained that "deepfake" combines "deep learning" and "fake," referring to its use in generating realistic photos, voices, videos, and other counterfeit products. This technology spans four key areas: text, voice, images, and videos. In defense applications, deepfakes can propagate false information extensively, leading to internal discord and undermining morale.
Huang cited a report from the Taiwan FactCheck Center (台灣事實查核中心) released on March 17, 2022, detailing how hackers released a deepfake video of Ukrainian President Zelensky delivering a surrender speech, aiming to undermine military morale. Addressing strategies for countering deepfakes, Huang noted that the NCSIST has the capability to assist the military in combating cognitive warfare using advanced technology.
MND spokesman Major General Sun Li-fang (孫立方) pointed out that during the Han Kuang exercises, misinformation often increases. Examples include false claims such as "the Han Kuang Exercise is meant to plan escape routes for the Commander-in-Chief" and "our military personnel might defect due to the heightened pressure." However, efforts by various government departments have reduced the spread and impact of such misinformation.
The MND stated that the military will proactively monitor public opinion and issue directions, intercepting misinformation before it spreads, regardless of whether it involves AI or deepfake technology.