AI Deepfake of President's Voice Could Violate 3 Laws, Taiwan Interior Ministry Says

Taiwan's Ministry of the Interior has reported that an individual accused of using AI to create a deepfake video of President Lai Ching-te's voice may have violated multiple laws. The case, involving a former director of the KMT's new media department, has drawn criticism from KMT legislators.

The ministry detailed that the police initiated an investigation on July 25 following a complaint. After an initial review of the video, which appeared to simulate the president's voice, the police notified the social media platform for review under its community guidelines. They also contacted the local precinct to verify if the account was impersonated and to determine the video's origin.

When officers visited the individual's home, they only left contact information for the family as the person was not present. The ministry stressed that no search was conducted, no one was detained, and no demand was made for immediate surrender. Subsequently, the individual contacted the police by phone, confirming they and their team created the account and video. The police then completed a preliminary identity check without requiring an immediate statement or forcing the video's removal.

The ministry clarified that police actions were legally based on the complaint, citing the Code of Criminal Procedure which mandates investigation and evidence preservation upon suspicion of a crime. They also noted that the use of AI to mimic voices or images could potentially breach laws on document forgery, defamation, and personal data protection, with final legal determination resting with judicial authorities. The police's notification to the social media platform followed its standard procedure, with content removal decisions left to the platform itself.