Recently, some people—including those with legal backgrounds—have once again been criticizing 'deepfakes,' claiming that any use of deepfakes constitutes a crime, especially when national leaders are involved, which they argue poses national security risks. In reality, deepfake technology 'itself' has never been the problem; the real issue lies in using this technology to commit illegal acts. Deepfakes are a neutral technology. For instance, using a celebrity's face to create a fake nude image raises concerns about defamation. However, using the same face to imagine a Western actress in the role of a Qing dynasty concubine—what harm is there in that?

Creating a video of 'Trump sexually harassing O' and spreading it as truth is problematic, but generating an image of 'Trump playing baseball with his daughter' to highlight their close relationship—is that illegal?

The key issue is whether false information is presented as real and causes actual harm to protected interests such as reputation or image rights. Deepfakes themselves are not the core problem. Even without using deepfake technology, fabricating text-based fake news or hiring look-alikes to stage events could equally constitute defamation. Publishing a false report claiming 'the president has declared martial law' could lead to charges of public intimidation or spreading rumors under public order laws—even without any AI involvement. Conversely, using deepfakes to enhance someone's image or fame might even be welcomed by the person depicted. In fact, many people routinely edit photos so heavily that they're unrecognizable—even to their own mothers—which is essentially a mild form of deepfake.