Skip to content
Submit incident
AII-1001-06441
ReportedReported by fewer than three independent publishers, with no official finding.Supporting sources are pending verification.

Chinese AI-Generated Propaganda Targets Taiwanese Social Media

Chinese content farms reportedly used AI to generate traditional Chinese articles specifically targeting Taiwanese social media users, with evidence including forgotten AI prompts identifying the targeting. Researcher Wang Hong-en presented multiple cases where the forgotten prompts revealed the deliberate, state-level nature of the propaganda campaign.

11 February 2026 Β· Chinese content farms Β· Misinformation and synthetic media
11 February 2026
AII-1001-06441
Quick facts
Category
Misinformation and synthetic media
Country
Taiwan
Date
11 February 2026
Drafted by
Raidu Incident Detection

What happened

Chinese content farms operating Facebook fan pages used AI to generate traditional Chinese articles targeting Taiwanese users, according to RAND researcher Wang Hong-en.

The AI-generated articles included forgotten AI prompts stating "targeting Taiwanese users, use traditional Chinese language, rewrite this article, control word count at 500, preserve the opening of the original, do not change the historical authenticity of the original."

Wang Hong-en identified multiple cases where prompt text was not deleted, and in some cases supplementary deletion attempts were made that left evidence in edit histories.

Reported impact

The articles promoted Chinese nationalism and anti-Japanese content aligned with Chinese official directions, while deprecating Taiwan, Japan, and the United States.

The content was part of broader cognitive warfare operations including promotion of Guo Tai-ming's candidacy and campaigns around pork imports.

Index analysis

Wang Hong-en noted that the targeted, cross-border nature with close contacts to Chinese official entities indicated deliberate state-level propaganda rather than individual editor oversight.

The AI-generated content made it increasingly difficult to detect cross-strait influence through analysis of traditional versus simplified Chinese character usage, requiring greater reliance on platform transparency and account geographic verification.

Classification

Category
Misinformation and synthetic media
Industry
Media, social platforms and marketing
Country
Taiwan
Harm status
Actual harm
Intent
Intentional
Timing
Post-deployment
Harm types
Public interest
Harmed
General public
Failure modes
Hallucination, Manipulation or dark pattern
Consequences
None recorded
AI systems
Unspecified AI text generation system

Sources and evidence

Sources not yet published

This record was researched and written by the Index from original reporting. Its sources are still being verified by an editor, and they are listed here once that check is complete. Until then the record carries no reviewer.

How sources are verifiedSuggest a source for this incident