Regulation of AI-generated disinformation by online platforms: A comparative analysis perspective

The rise of Generative artificial intelligence has transformed the generation and dissemination of disinformation, accelerating its spread and expanding its reach. This thesis will examine how the European Union, the United States, and China are addressing the regulation of AI-generated disinformation and placing primary regulatory liability on online platforms. Using comparative study, case study, and doctrinal study, this thesis will analyze the historical evolution of platform liability regimes and current regulatory measures across jurisdictions. The findings reveal that the EU’s regulatory approach emphasizes transparency and requirements under the Digital Services Act and content moderation obligations imposed on VLOPs, while the US’s regulatory framework prioritizes free speech and provides intermediaries with immunity under Section 230. China, on the other hand, adopts a state-led regulatory approach that encourages online platforms to proactively conduct content moderation. The significance of this research lies in analyzing how inadequacies or ambiguities in various jurisdictions’ laws and regulations lead to enforcement difficulties and how malicious disinformation producers exploit these inadequacies to circumvent regulation. Furthermore, it offers feasible recommendations for establishing a cross-jurisdictional collaborative framework for addressing AI-generated disinformation.