Chinese paper outlines algorithmic cognitive warfare
- The Special Competitive Studies Project published Libby Lange’s paper tracing how Chinese military writers describe “algorithmic cognitive warfare” as AI-driven influence operations. - Lange says the concept combines recommendation systems, large language models, deepfakes, and personal-data analysis to steer beliefs at granular scale in real time. - The paper lands amid wider U.S. and UK debates over AI security, deterrence, and defense readiness. (scsp.ai)
Cognitive warfare means trying to change how people think, feel, and decide. A new paper from the Special Competitive Studies Project says Chinese military writers increasingly frame artificial intelligence as the engine for doing that at scale. (scsp.ai) The paper, published by SCSP and written by Libby Lange, examines Chinese scholarly and military-affiliated writing on what it calls “algorithmic cognitive warfare.” It says the idea centers on using algorithms to target, influence, and divide foreign populations. (scsp.ai 1) (scsp.ai 2) In plain terms, the model works like a recommendation feed crossed with a political influence campaign. Algorithms sort people by behavior and emotion, then push tailored text, video, or social content meant to shape beliefs or trigger reactions. (scsp.ai) Lange writes that Chinese authors describe two key layers: artificial intelligence systems that analyze and track individuals, and platform recommendation systems that decide what content people see. The paper says that combination could turn mass persuasion into individualized persuasion. (scsp.ai) The report says success in that model depends on detailed personal data. It argues that Chinese military and intelligence services would need large volumes of behavioral and demographic data to deliver the level of tailored messaging those writings envision. (scsp.ai) The paper also places the idea inside a longer Chinese discussion of the “cognitive domain,” which it says appeared in Chinese scholarship by 2010. In that framework, the target is not just troops on a battlefield but civilian populations, political will, and public trust. (scsp.ai) Chinese state and military-linked commentary outside the SCSP paper shows similar themes. A March 2023 article reposted by a Tibet regional cyberspace office described cognitive warfare as a fight over trust and pointed to deepfakes, social bots, and machine-learning systems that can detect, classify, and exploit online narratives. (xzdw.gov.cn) A January 22, 2026 article on China Military Online described “cognitive electronic warfare” as a system that can sense threats, make decisions, evaluate effects, and update its knowledge base in a closed loop. That article focused on electromagnetic warfare, but it used the same language of autonomy, adaptation, and machine-led response. (81.mil.cn) SCSP says there is still a gap between doctrine and real capability. The paper does not claim the People’s Liberation Army can already execute the full model it describes, but it argues the writings offer a direct view into how some Chinese strategists think about future conflict. (scsp.ai) The backdrop is a broader Western security debate over artificial intelligence in defense and intelligence. The U.K. Parliament’s POST office warned in January 2025 that AI, disinformation, and cyber tools are becoming a growing national security issue, while the U.K. government’s AI Opportunities Action Plan called for secure AI infrastructure and collaboration with the intelligence community. (post.parliament.uk) (gov.uk) In Washington, recent Pentagon and congressional documents have cast China as the main pacing challenge and put more emphasis on AI-enabled warfare, autonomy, cyber, cognitive, and electronic conflict. That is the setting in which Lange’s paper is being read: less as a prediction of one weapon, and more as a map of how influence operations could be automated. (media.defense.gov 1) (media.defense.gov 2) The thread running through all of it is simple: if platforms can learn what holds attention, military planners will ask whether those same systems can learn what changes minds. Lange’s paper says Chinese writers are already exploring that question in detail. (scsp.ai)