Skip to content
Submit incident
AII-1001-58013
ReportedReported by fewer than three independent publishers, with no official finding.

China uses AI-generated personas to interfere with Taiwan elections

Internal documents from a Beijing-based company allegedly describe AI systems capable of generating synthetic personas that mimic influential people's dialects and voices to spread information and manipulate public opinion in Taiwan ahead of its 2026 and 2028 elections, according to Japanese media reporting on materials obtained by an American university.

27 December 2025 · Government of China · Misinformation and synthetic media
27 December 2025
AII-1001-580132 sources ↓
Quick facts
Category
Misinformation and synthetic media
Country
Taiwan
Source
Yahoo News (Taiwan)
Date
27 December 2025
Drafted by
Raidu Incident Detection

What happened

American Vanderbilt University's National Security Institute obtained an internal document from Beijing-based company 中科天璣 (Zhongke Tianji) in 2025, which detailed its methods for generating AI synthetic personas to influence Taiwan's politics. The 400-page document described techniques to monitor online discourse, identify influential opinion leaders, analyze their psychological traits, values and dialects, then generate synthetic personas mimicking their voice and speaking patterns to engage in dialogue and shift their views.

The system was designed to target topics of high concern to the Chinese Communist Party central leadership, including Hong Kong, Taiwan, and US-China relations. The company reportedly claimed to provide 'situational awareness, analysis and opinion guidance technology' for these issues. When asked, 中科天璣 denied the New York Times' reporting on the document.

Reported impact

A Taiwan AI industry representative stated that with rapid advances in generative AI, fake accounts on social media have become nearly impossible to distinguish, posing a serious challenge to democratic society. External observers believe China aims to prevent incumbent President Lai Ching-te from winning re-election in 2028, with the 2026 local elections serving as a preliminary test of strength for the subsequent opinion warfare.

Index analysis

The system monitored online spaces to identify influential opinion leaders, then analyzed their psychological traits, values and dialect, and generated AI-generated synthetic personas with high similarity to them. These synthetic accounts could send messages proactively and conduct natural conversation to gradually influence the target's cognition and stance.

Governance controls

Taiwan's Mainland Affairs Council held a consultative meeting in early December 2025 to warn of China's opinion warfare, proposing measures to strengthen responses to false information and deepen civic education.

Classification

Category
Misinformation and synthetic media
Industry
Media, social platforms and marketing
Country
Taiwan
Harm status
Potential harm
Intent
Intentional
Timing
Unknown
Harm types
Public interest, Reputational
Harmed
General public, Government
Failure modes
Voice clone, Manipulation or dark pattern
Consequences
None recorded
AI systems
AI persona generation and opinion manipulation system (中科天璣)

Sources and evidence

This record was researched and written by the Index, based on the following original reporting.

Yahoo News (Taiwan)
《讀賣新聞》揭中國偽造「AI台灣腔」企圖介選2026、2028
27 December 2025
tw.news.yahoo.com
China uses AI-generated personas to interfere with Taiwan elections
Sep 2026AI-powered influence campaigns manipulate social media in US and IsraelConfirmedMisinformation and synthetic mediaAnadolu AjansıThe Jerusalem PostMar 2026EU Report Reveals China's Use of AI for Disinformation and HarassmentConfirmedMisinformation and synthetic mediaLiberty Times NetUDNCentral News Agency+2Jan 2026China Creates AI-Generated Deepfake Video of Taiwan Students Supporting UnificationReportedMisinformation and synthetic media新頭殼 NewtalkNOWnews 今日新聞Aug 2025Chinese AI company targeted 2024 Taiwan election and Hong Kong dissent with bot networksReportedMisinformation and synthetic mediaNew York Times (Chinese)Liberty Times NetJul 2025China: AI deepfakes and data leaks pose national security risksConfirmedMisinformation and synthetic media中华网军事频道news.cctv.com