AI Is Lowering the Barriers to China’s Information Warfare Against Japan

Japan's Ministry of Defense published its 2026 Defense of Japan white paper, highlighting China's rapid advancements in AI and its potential ramifications for regional security. Despite these developments, the immediate threat to Japan may stem from China's use of generative AI to facilitate information warfare targeting Japanese society.

Prime Minister Sanae Takaichi made comments in the Diet in late 2025, linking a Taiwan contingency to Japan's security. Since then, disinformation targeting Japanese political figures and domestic issues has increased in both quantity and quality. Research indicates that China-linked influence operators are leveraging AI to tailor their activities to Japan's online environment and enhance the efficiency of their operations.

The linguistic and cultural barriers once hindering Chinese information operations against Japan are diminishing. Content that previously contained grammatical errors, awkward translations, and unnatural expressions is now harder for ordinary users to detect. If such capabilities are employed to sway elections, discredit politicians, or exacerbate social divisions, the consequences could extend beyond individual false narratives.

A February report by Nikkei revealed that over 400 China-linked social media accounts disseminated negative content about Takaichi during Japan's House of Representatives election. Additionally, the Yomiuri Shimbun, citing research by Japan Nexus Intelligence, reported that thousands of coordinated accounts on X were involved in sharing and amplifying related content.

These narratives aimed not only to attack Takaichi personally but also to associate Japan's defense expansion with concerns over social welfare, suggesting that military growth disadvantaged ordinary citizens.

Historically, these accounts often displayed telltale signs, such as recently created accounts with Japanese reflecting Chinese linguistic habits or AI-generated images containing visible errors. However, the sophistication of these operations is evolving. The Sankei Shimbun's tracking of PAPERWALL, a China-linked information network consisting of 15 websites posing as Japanese news outlets, suggests an increasingly localized approach to influence operations.

OpenAI's February report further sheds light on the integration of AI in China-linked influence campaigns targeting Takaichi. While the nationality of the AI model is less significant, the key takeaway is that generative AI is reducing the cost of cross-language information operations. Large language models can now assist operators in correcting grammar, adjusting tone, generating images, and producing multiple narrative versions for distinct audiences within minutes.

Japan's linguistic and cultural context remains a significant advantage, as Chinese operators may improve translations and content generation but struggle to grasp the nuanced linguistic and political differences among various Japanese demographics.