我國深度偽造技術(Deepfake)之治理與修法評析

Analysis of the Legal Framework about Deepfake Technology in Taiwan

臺灣於2021年底爆發了「網紅換臉色情片群組」事件，使深度偽造技術濫用議題受到空前的關注。法務部與行政院亦於2021年底與2022年3月，公布深度偽造影像的管制與處罰的修法方向與刑法修正草案，且該草案於2023年1月7日經立法院三讀通過。
本文將深度偽造技術的潛在危害分成兩大面向：(1)涉及數位性暴力議題的「性影像」相關Deepfakes、(2)涉及假訊息管制問題的「傳遞不實訊息」Deepfakes。並以兩個面向為核心，比較與分析外國與我國現行法、新修法、以及未來可能管制深度偽造影像的手段。

Deepfake is a technology that uses artificial intelligence (machine deep learning) to produce fake but almost realistic images and videos. Recently, with the continuous development of such technology, countries around the world began to notice the potential harm through the misuse of deepfake technology. At the end of 2021, the“Deepfake sex tape chatroom scandal”attracted the attention of society. Lawmakers in Taiwan were asked to develop relevant penalties and regulations toward such deepfake materials. The Taiwan Ministry of Justice submitted a bill targeting deepfakes to the Executive Yuan in November 2021 and the Executive Yuan further announce a bill to amend the current Criminal Code in March 2022. That bill was approved by the Legislative Yuan and became a part of the Criminal Code
The potential harms of deepfakes could be divided into two main categories: those deepfake content related to“sex”and those related to“misinformation”. Based on these two categories, this article will introduce the current development of deepfake regulations in the United States, Korea, China, and the Europe Union. Furthermore, this article analyzes the current law and the new bill presented by the Taiwan Ministry of Justice submitted and Executive Yuan in a comparative law aspect. Finally, this article will discuss the future of Taiwan’s deepfake regulations by presenting the possibilities and importance of establishing platform responsibilities and technological ethics.