論假訊息之規管與治理

假訊息泛指造成「惡、假、害」的不實資訊，其中對於「假」之定義，並不必然要全部為假才構成之；事實上，半真半假的假訊息，更是難於查證而利於誘導民眾受騙。鑑於社會整體資訊生態系統已然失調，假訊息生成後，難以被快速查證下架，惟更多的假訊息又持續生成，甚至在查證單位出面闢謠後，基於逆火效應的偏誤，仍有多數人信奉假訊息為真諦，而社群網站假訊息被轉傳次數遠高於澄清訊息的數倍現象，係假訊息難被根除的真實寫照。是以，假訊息肆虐防不勝防，各國陸續浮出欲為規管聲浪，有著政府應著手介入規管之呼籲。本文擬從法釋義學角度直搗言論自由理論核心，首要釐清國家得否祭出法律對言論自由進行規管的課題，並梳理我國違憲審查程序之合憲性事由，將用以檢視外國與我國為解決假訊息肆虐所設法制之利弊及合憲性，助於文末規範途徑之選擇。從而，本文提出國際上廣為採用的三種防制假訊息策略，分別為：「行為人規管模式」、「中介者規管與治理模式」，及「事實查核治理模式」，並帶入我國相關法律及機制進行檢驗。最終，本文認為，融合傳統規管思維搭配創新監理模式的中介者規管與治理模式，相形之下，係對言論自由侵害較小且具備執行成效的立法模式。

關鍵字

言論自由；假訊息；社群媒體；網路中介者；規管模式；網路治理；事實查核機制

並列摘要

Disinformation generally refers to untrue information that causes "evil, falsity, and harm." The definition of "fake" does not necessarily mean that all of the information is false; in fact, half-true, half-false disinformation is more difficult to verify and is conducive for deception. Since the overall "information ecosystem" of society is already dysfunctional, it is difficult to quickly verify and remove fake messages once they have been generated, eventually creating even more fake messages. Even after the verification unit has been deployed to dispel a rumor, due to the backfire effect, most people will still believe that fake messages are true. The phenomenon that fake messages on social networking sites are forwarded several times more often than clarification messages truly reflects the difficulty of eradicating fake messages. As disinformation is rampant and difficult to prevent, increasingly many voices from various countries are calling for government intervention and regulation. This article intends to cut straight to the core of the theory of freedom of speech from the perspective of legal interpretation. First, it will clarify whether the state can regulate freedom of speech, then sort out the issue of the constitutionality of Taiwan's unconstitutional review process to examine the advantages, disadvantages, and constitutionality of the legal system established by foreign countries and Taiwan to deal with the disinformation rampage, and finally offer suggestions of normative ways at the end of the article. This article summarizes three widely adopted strategies that seek to prevent disinformation in the world, namely the "actor regulatory model," "intermediary regulatory and governance model," and "fact-checking governance model" as examined with reference to Taiwan’s relevant laws and mechanisms. This article concludes that the second model which combines traditional regulation thinking with the innovative supervision model, is a legislative model that is comparatively less infringing on freedom of speech yet also effective in implementation.