Taiwan Bolsters AI Defenses Amid Fears of Chinese Election Interference

Taiwan is strengthening its defenses against artificial intelligence-driven election interference as analysts warn that China could increasingly use AI tools to influence the island's November 28 local elections.

Taiwan's ruling Democratic Progressive Party (DPP) has warned that Beijing could deploy artificial intelligence to target specific groups of voters, create convincing fake social media identities and spread fabricated audio, video or damaging allegations in the final days before voters head to the polls.

The concerns have prompted Taipei to intensify efforts to detect and counter AI-generated manipulation, regardless of whether such activity originates domestically or abroad.

China claims Taiwan as part of its territory, while Taiwan rejects Beijing's sovereignty claim and maintains that the island's future should be determined by its people.

Researchers say AI gives influence operations new capabilities, allowing campaigns to generate content rapidly, operate large networks of fake accounts and customize political messages for particular groups of voters.

Sze-Fung Lee, an independent researcher specializing in Chinese hybrid warfare, said AI was likely to play a greater role in the upcoming elections.

She said previous Taiwanese elections had already seen the use of AI-related techniques, including deepfakes used in defamation campaigns, coordinated activity by fake accounts and increasingly sophisticated efforts to target individual voters.

William Yang, a Northeast Asia analyst with the International Crisis Group, also said Taiwan had legitimate reasons to be concerned about Chinese interference supported by advanced AI technology.

China has a long history of alleged interference in Taiwan's elections and, according to analysts, AI could significantly expand the scale and sophistication of those efforts.

Beijing has repeatedly denied interfering in Taiwan's electoral process. Chinese authorities accuse the DPP of pursuing separatist policies and exploiting tensions across the Taiwan Strait for political advantage.

Difficult Election for Taiwan's Ruling Party

The DPP's warning comes as it faces a challenging local election campaign.

The opposition Kuomintang (KMT) won 14 mayoral and county chief positions in the 2022 local elections, compared with five for the DPP. The KMT has also held a majority of those local leadership positions during the past two election cycles.

Yang said the KMT was currently expected to win a larger number of counties, although the DPP could remain competitive in several contests.

The ruling party is expected to closely monitor disinformation campaigns that could damage its electoral prospects and investigate whether suspicious activity can be linked to Chinese actors.

The KMT favors greater engagement with Beijing but rejects claims that it is a pro-China party. The DPP, by contrast, has sought closer relations with countries including the United States and Japan.

In February, the KMT said it opposed any interference in Taiwan's elections by outside actors.

Despite the deep divisions between the two parties over relations with Beijing, local elections in Taiwan are typically dominated by domestic concerns.

The DPP placed a strong emphasis on national security and resistance to Chinese pressure during the 2022 vote, but suffered one of its worst local election performances in years. The result prompted then-President Tsai Ing-wen to step down as party chair.

Lessons From the 2024 Presidential Election

Taiwan's experience during its 2024 presidential election has heightened concerns over the potential use of AI in political influence campaigns.

The DPP said AI-generated material circulating during that election included videos promoting an electronic book titled "The Secret History of Tsai Ing-wen", which contained allegations involving the former president. Authorities also investigated a suspected fabricated recording attributed to businessman and former presidential candidate Terry Gou.

Research by Taiwan-based civil society organization Doublethink Lab found that generative AI was used during the 2024 election to support large-scale spam operations and enable inauthentic accounts to produce more varied posts and alter their publishing schedules.

The research also found a shift in the apparent nature of influence activity. Earlier campaigns had relied more heavily on amplifying Chinese state media and prominent accounts on Chinese social media platforms, while much of the activity identified during the 2024 election originated from unofficial social media accounts.

However, researchers cautioned that establishing who is behind such operations remains difficult.

The existence of AI-generated material alone does not prove that China created or coordinated it. Investigators must instead examine account networks, patterns of online behavior and links to state-backed media or other known actors.

AI could make that process even more challenging by enabling influence operators to imitate Taiwanese language and expressions, engage with local political issues and make their online activity appear more authentic.

A February report from the Leiden Asia Centre similarly warned that AI could make disinformation significantly cheaper and easier to produce on a large scale.

The report nevertheless found that the effectiveness of political manipulation in Taiwan depended heavily on existing political divisions, social grievances and social media platforms that tend to reward emotionally charged material.

Rather than creating divisions from scratch, AI may therefore be more effective at magnifying political tensions that already exist.

Taiwan Strengthens Election Safeguards

Taiwan's Central Election Commission is preparing additional measures to counter AI-generated election manipulation.

The commission is establishing a specialized task force and an online reporting platform that will allow candidates to flag suspected deepfakes.

Officials will assess reported material and, where cases are confirmed, refer them to social media and other online platforms for potential action.

The system could prove particularly useful in cases involving fabricated audio or video attributed to identifiable political candidates.

However, broader influence operations could prove considerably more difficult to combat. Individual users can post political opinions, misleading claims or manipulated material without necessarily forming part of a clearly identifiable coordinated network.

Closed online communities are another major challenge.

Political content shared through private or semi-private groups on messaging platforms such as LINE can be difficult for election officials, researchers and fact-checkers to monitor.

Taiwan's preparations are also taking place against a backdrop of persistent concerns over Chinese cyber activity.

Taiwan's National Security Bureau reported that critical infrastructure faced an average of 2.63 million cyberattacks originating from China each day during 2025. That represented a 6% increase from the previous year and a 113% increase compared with 2023.

According to the bureau, some attacks occurred around Chinese military exercises or other politically significant events in Taiwan.

The cyberattack figures do not establish that AI was involved in election interference. However, researchers say they reinforce concerns about the scale of the broader pressure Taiwan faces from China.

Lee said the government's warnings should not be dismissed as excessive caution, pointing to the frequency of cyberattacks and what she described as continuing cognitive warfare directed at Taiwan.

With AI technology becoming increasingly sophisticated and accessible, Taiwan is now preparing for an election environment in which fabricated content can be produced faster, distributed more widely and made increasingly difficult to distinguish from genuine political material.