A short political promotional video released on July 24, which allegedly used artificial intelligence (AI) to mimic the voice of President William Lai (賴清德) in a criticism of the government’s handling of a cooking oil safety scandal, has become the focus of political debate across party lines, highlighting the controversial issue of regulating AI-generated deepfake content.
The video was released a day before the Chinese Nationalist Party (KMT) held a rally on Ketagalan Boulevard over the cooking oil scandal.
Shortly after the video was released, former KMT Youth Department director Wei Chun-yu (韋淳祐), who created the video campaign, expressed frustration at being approached by the Criminal Investigation Bureau, calling their questioning an infringement on the freedom of expression.
The Democratic Progressive Party (DPP) condemned the video, saying that the use of deepfake content could mislead the public into believing it was a genuine statement from the president. Legislators called for a judicial investigation, citing potential violations of the Personal Data Protection Act (個人資料保護法) and infringement of the personal rights of the president.
Taipei Mayor Chiang Wan-an (蔣萬安), who called for the protest, responded that the judicial system has become a tool for political persecution, and added that the police questioning Wei constituted state intimidation.
Minister of the Interior Liu Shyh-fang (劉世芳) said that while Taiwan protects freedom of speech, AI-generated impersonation is not protected by this right and that the police should dutifully and lawfully investigate when deepfakes are illegally used to spread misinformation and disrupt public order. The Ministry of the Interior said that the police have a duty to respond to a report, and Wei’s personal freedoms were not restricted as a result of the questioning.
DPP Legislator and Taipei mayoral candidate Puma Shen (沈伯洋), also an associate professor of criminology, last week said using deepfakes to impersonate a country’s leader is a high-risk action typically associated with hostile state actors, especially given Taiwan’s geopolitical situation. Regulation regarding deepfakes is widely limited to fraud, making this a special case, he said, adding that standalone legislation is needed.
The Public Officials Election and Recall Act (公職人員選舉罷免法) has deepfake provisions for election and recall misinformation, while other cases rely on general criminal, civil and personal-data laws.
The case highlights the issue of whether the Artificial Intelligence Basic Act (人工智慧基本法), which took effect in January, can keep up with emerging technology, and if a new law with clearer legal liabilities and penalty mechanisms should be enacted to explicitly regulate AI practices.
Whether an AI disclaimer would be sufficient to avoid legal liability needs to be further discussed, as well as if harsher penalties and strict criminal regulations should be applied to cases in which public figures, political candidates and government officials become the targets of deepfakes due to the potential for public harm, such as destabilizing democratic elections, fueling public panic or even damaging national security.
Academics and law experts have warned that the spread of deepfake content, especially the impersonation of government officials, could undermine democratic decisionmaking by distorting public perception and eroding trust in facts, with some citing a deepfake video showing Ukrainian President Volodymyr Zelenskiy seemingly telling soldiers to surrender to Russian aggression in March 2022.
The European Parliament in 2024 passed the world’s first comprehensive legal framework regulating AI — the EU AI Act — which was entered into force on Aug. 1, 2024, with provisions going into effect in stages. Major transparency rules, mandatory deepfake labeling and enforcement powers took effect yesterday.
In the US, the Take It Down Act passed in May last year banned the publication of nonconsensual sexual imagery, while many states have laws requiring clear disclaimers or watermarks on AI-generated content in advertisements or elections.
Meanwhile, a Japanese Ministry of Justice expert panel on Monday last week concluded a study of AI by broadly approving a draft report about extending legal protections to human voices against AI-generated cloning and misuse.
As Lai has championed his vision of turning Taiwan into an “AI island,” maybe it is time for Taiwan to seriously review and update its laws on AI regulation to keep up with rapid technology changes.
Over the past few months, Honduras has quietly reduced its diplomatic presence in China. Former Honduran ambassador to China Salvador Moncada left his post on Jan. 26, one day before Honduran President Nasry Asfura took office, and no replacement has been announced ever since. The consulate and the embassy have remained largely inactive for months, reports said. The downgrade process culminated this month when Honduras officially announced the closure of its consulate in Shanghai. By itself, the closure of a consulate does not necessarily indicate a geopolitical realignment. Governments regularly reorganize diplomatic networks for budgetary reasons. Yet, when viewed alongside
A lecture by Tao Yi-fen (陶儀芬), an associate professor in National Taiwan University’s Department of Political Science, offered an insightful assessment of US-China strategic competition and Taiwan’s future. Drawing on academic research and her national security experience, Tao argued that China faces mounting economic and political constraints, that “the East rising and the West declining” is a false narrative, and that time remains on the side of the US and Taiwan. Her analysis deserves serious consideration. Yet it also illustrates an important point: Theories of international relations often produce different conclusions from the same set of facts. As a specialist in international
RSS