Japanese AI startup Sakana AI, in collaboration with the Yomiuri Shimbun, has unveiled a proprietary system designed to visualize the reality of "cognitive warfare" across vast social media datasets. Moving far beyond simple keyword aggregation, this initiative deeply parses the context and nuances behind posts to expose state-level propaganda and organized information maneuvers. By analyzing 1.1 million social media entries, the system successfully identified specific narrative frameworks used in Chinese criticism of Japan. This technology excels at detecting subtle signals within the massive information space that were previously overlooked.

The technical foundation of this system is a method called "Novelty Search," which coordinates three distinct language models. This "collective intelligence" approach allows multiple models to layer inferences from different perspectives, enabling high-granularity information discovery and hierarchical organization that was difficult to achieve with conventional methods. A key feature is its ability to extract latent narratives—such as "intervention in Taiwan issues"—even from posts that lack direct keywords. Furthermore, the system autonomously generates hypotheses based on extracted data and presents the reasoning and evidence to experts, establishing a "human-in-the-loop" process that significantly enhances human analytical efficiency.

In practice, the system derived hypotheses regarding the processes through which criticism campaigns were triggered by the remarks of Japanese politicians, and the validity of these insights was subsequently verified by Yomiuri Shimbun’s investigative reporting. This marks a new form of intelligence where AI discovers patterns and insights invisible to humans, which then serve as the basis for high-level human decision-making. Sakana AI has explicitly stated that "Defense and Intelligence" will be its most critical focus area alongside finance, signaling a clear commitment to contributing to national security through advanced Japanese technology. AI-driven narrative analysis is poised to become an indispensable tool for predicting geopolitical risks in the future.

Japanese AI startup Sakana AI, in collaboration with the Yomiuri Shimbun, has unveiled a proprietary system designed to visualize the reality of "cognitive warfare" across vast social media datasets. Moving far beyond simple keyword aggregation, this initiative deeply parses the context and nuances behind posts to expose state-level propaganda and organized information maneuvers. By analyzing 1.1 million social media entries, the system successfully identified specific narrative frameworks used in Chinese criticism of Japan. This technology excels at detecting subtle signals within the massive information space that were previously overlooked.

The technical foundation of this system is a method called "Novelty Search," which coordinates three distinct language models. This "collective intelligence" approach allows multiple models to layer inferences from different perspectives, enabling high-granularity information discovery and hierarchical organization that was difficult to achieve with conventional methods. A key feature is its ability to extract latent narratives—such as "intervention in Taiwan issues"—even from posts that lack direct keywords. Furthermore, the system autonomously generates hypotheses based on extracted data and presents the reasoning and evidence to experts, establishing a "human-in-the-loop" process that significantly enhances human analytical efficiency.

In practice, the system derived hypotheses regarding the processes through which criticism campaigns were triggered by the remarks of Japanese politicians, and the validity of these insights was subsequently verified by Yomiuri Shimbun’s investigative reporting. This marks a new form of intelligence where AI discovers patterns and insights invisible to humans, which then serve as the basis for high-level human decision-making. Sakana AI has explicitly stated that "Defense and Intelligence" will be its most critical focus area alongside finance, signaling a clear commitment to contributing to national security through advanced Japanese technology. AI-driven narrative analysis is poised to become an indispensable tool for predicting geopolitical risks in the future.