論深度偽造影音影像之規制：以散布之處罰為中心

Abstract

深度偽造影音影像，作為人工智慧的新興應用，其真假難辨，撼動人類既有認知。自迄今商業、教育與醫療等用途以觀，前景不可限量；然與此同時，所帶來之危害也不容小覷。近年來，世界各國無不投身於思索因應之道，我國亦不例外。經社會事件帶動輿論，我國推動深度偽造影音影像之立法。除針對色情深度偽造影音影像與政治深度偽造影音影像已增訂新法外，關於不同類型深度偽造影音影像之一般性處罰，其立法芻議亦如火如荼地展開商討。本論文嘗試自深度偽造影音影像之源起、發展、定義與特徵以釐清其危害，盤點我國既有法規範因應此等危害是否有所不足，並借鏡各國立法動向，探討我國增訂新法與立法芻議是否妥適，進而提出本論文立法建議。期盼檢討之結果，對於因應深度偽造影音影像之法制建構，有所助益。

Deepfakes are emerging applications of artificial intelligence that are so indistinguishable from the reality that they are challenging human perceptions. Since use in business, education and medication, their future is unlimited, but the harm they bring is not to be underestimated. In recent years, many countries have been engaged in considering ways to address the issue, and our country is no exception. As a result of social events and public opinion, Taiwan has been promoting legislation on deepfakes. In addition to the enactment of new laws on pornographic deepfakes and political deepfakes, preliminary discussions on the legislation of different types of deepfakes are also in progress. This paper attempts to clarify the harm deepfakes bring with their origins, development, definitions and characteristics, to investigate the inadequacy of our current laws and regulations in response to these harms, to examine the appropriateness of the new laws and legislative regurgitations in our country with the legislative trends in other countries, and to make legislative recommendations. Hoping this paper may help to build the construction of a legal framework to deal with deepfakes.