Taipei Mayor: 'AI Deepfake' of Lai Ching-te Not Illegal; Top Prosecutor's Office Warned Twice of Dangers
Translated from Chinese, summarized and contextualized by DistantNews.
At a glance
- Taipei Mayor Chiang Wan-an confirmed his team produced an "AI deepfake" promotional video featuring President Lai Ching-te's voice, claiming it does not constitute a crime.
- The video was created for a "725 anti-drug oil" rally and has drawn criticism for potentially misleading the public.
- Taiwan's Supreme Prosecutors Office has previously warned about the dangers of generative AI and deepfakes in elections, urging caution.
Taipei Mayor Chiang Wan-an has confirmed that his team commissioned the creation of an "AI deepfake" promotional video that synthesizes President Lai Ching-te's voice. Chiang maintains that the video, produced for a "725 anti-drug oil" rally, does not violate the law because it was clearly labeled as AI-generated.
However, this assertion faces scrutiny, as Taiwan's Supreme Prosecutors Office has twice this year flagged generative AI and deepfakes as significant threats to democratic processes during elections. Legal experts argue that local officials should exercise greater caution in their public statements and avoid downplaying the risks associated with deepfake technology.
The controversy centers on a promotional video for the rally, created by Wei Chun-yu, an administrator for the "Principal Lai" fan page and former director of the KMT's new media department. Wei is reportedly under investigation for allegedly using AI to fabricate President Lai's voice. Chiang's defense that the video's AI origin negates its criminality has been met with criticism, with some suggesting it could create a chilling effect on investigations.
Legal professionals familiar with election law emphasize that the judiciary views deepfakes with high concern. The issue is not about restricting political commentary but about the potential for highly realistic AI-generated audio and video to mislead audiences into believing individuals said things they never did. This fundamentally differs from traditional forms of imitation or commentary.
The core legal question, according to prosecutors, lies in whether AI-generated content, which makes individuals appear to say things they haven't, can mislead the public, influence judgment, or disrupt democratic procedures. While political satire is a consideration, the authenticity of the generated content and its potential impact on public discourse are paramount. Cases will be judged based on the video's content, production methods, distribution, and actual effects, rather than simply labeling it as satire.
Originally published by Liberty Times in Chinese. Translated, summarized, and contextualized by our editorial team with added local perspective. Read our editorial standards.