Abstract:  The rapid development of artificial intelligence generated content (AIGC) has brought unprecedented and revolutionary experience for human production and life, while promoting the continuous innovation of the AI industry, it is also necessary to face up to the various types of risks arising from this technology. In the past two years, the amount of deepfake content driven by generative technologies has been increasing, and the audiovisual manipulation of the public by technological manipulators through false information has gradually become a severe challenge that needs to be addressed in today’s society. Based on the theory of information manipulation, a comprehensive theoretical framework covering “motive-process-result” is constructed. Under this framework, a detailed and in-depth discussion is made on the key dimensions of manipulators’ motives, means, communication channels, audience weaknesses and real-world impacts of audiovisual manipulation by means of deepfake technology, and the internal mechanisms of the whole process of deepfake has been comprehensively revealed. On this basis, targeted governance strategies from three aspects of agile governance, technological innovation and global co-governance are proposed.

Key words: artificial intelligence generated content, deepfake, internal mechanisms, governance strategies