Hi Gemini, I just read something in the South China Morning Post about Algorithmic Warfare, a reseach paper by Students of the National Defense University´s Joing Operations College. - Do you know something about it?
Yes, I'm quite familiar with this thread of research! The South China Morning Post (SCMP) frequently tracks papers coming out of China’s top military academies, and the specific focus you are referring to centers heavily on a concept called "Algorithmic Cognitive Warfare" (算法认知战).
SCSP | Substack
When students and researchers at the PLA National Defense University (NDU)—specifically within the Joint Operations College—write about this, they are looking at how to merge advanced AI with the psychological and electronic domains.
The core concepts driving these NDU research papers generally focus on the following pillars:
1. Moving Beyond "Informatized" to "Intelligentized" Warfare
For years, the Chinese military focused on informatized warfare (networking forces via IT, digital communications, and satellites). The current doctrine has shifted to intelligentized (智能化) warfare. NDU researchers argue that tomorrow's conflicts won't just be won by who has the fastest network, but by who has the superior algorithms to compress the OODA Loop (Observe, Orient, Decide, Act) to near-instantaneous speeds.
NDU Press - National Defense University+ 2
2. Targeting the "Cognitive Domain"
The Joint Operations College papers frequently treat the human mind as a distinct battlefield, separate from land, air, or sea. They focus on how AI recommendation algorithms (similar to the logic that drives commercial apps like TikTok) can be weaponized during a crisis. The goal is to:
Inject highly tailored, micro-targeted misinformation or narrative shifts to fracture an adversary's political resolve or confuse public perception.
SCSP | Substack
Paralyze an opponent's decision-making timeline before physical kinetic actions even begin.
3. Automated Command & Control Support
Interestingly, Western defense analysts tracking these NDU papers note a specific internal vulnerability China is trying to solve: a historical lack of recent combat experience and a highly centralized, rigid command structure. NDU students heavily research AI decision-support systems to act as "cognitive co-pilots." The idea is to use algorithms to ingest massive amounts of open-source and sensor data, instantly offering theater commanders optimized tactical choices to reduce human hesitation.
National Defense Magazine
Since we previously touched on "Hyperwar" and Electronic Warfare (EW), you can think of algorithmic warfare as the brains driving that entire ecosystem. In a Hyperwar scenario, automated EW systems and algorithmic cognitive operations trigger simultaneously—jamming physical signals while algorithmically flooding the information space to create total tactical blindness.
Are you looking at this from the perspective of how the West is trying to counter these algorithmic techniques, or are you more curious about the specific AI tools (like large language models or deepfakes) the paper mentions using?
is there any way the West is able to counter algorithmic chinese efforts?
Countering this algorithmic and cognitive push from China is currently one of the highest priorities for Western defense networks. However, because Western democratic societies are inherently open, they face a structural vulnerability that a tightly controlled information ecosystem like China’s does not.
Verfassungsblog
To bridge this gap, NATO and the U.S. Department of Defense (DoD) have shifted their approach from traditional "counter-propaganda" to active, data-driven defense. The strategy relies on a multi-layered defense model:
┌─────────────────────────────────────────┐│1. THE ELECTROMAGNETIC LAYER │ -> Jamming data relays & blinding ISR sensors└────────────────────┬────────────────────┘▼┌─────────────────────────────────────────┐│2. MACHINE-LEARNING DEFENSE │ -> Poisoning training sets & model-weight hacking└────────────────────┬────────────────────┘▼┌─────────────────────────────────────────┐│3. THE HUMAN/COGNITIVE DOMAIN │ -> "Narrative Intelligence" & active attribution└─────────────────────────────────────────┘
1. Striking the "Algorithmic Fuel" (The Electronic & Cyber Layer)
Algorithms cannot function without clean data, computing power, and reliable transmission. Western strategy increasingly focuses on degrading the infrastructure that Chinese algorithms require to make decisions:
Sensor Blinding: Western electronic warfare (EW) units focus heavily on techniques to spoof or blind China's Intelligence, Surveillance, and Reconnaissance (ISR) platforms. By feeding Chinese battlefield algorithms corrupted, chaotic, or non-patternable sensor data, the AI's tactical recommendations break down.
Data Poisoning & Model Hacking: In the cyber domain, Western offensive operations target the machine learning pipelines themselves. If an adversary's military AI is trained on data sets that have been subtly altered ("poisoned") by Western cyber commands, the algorithm will draw fatally flawed conclusions during a crisis.
2. "Narrative Intelligence" and Early Detection
In the cognitive domain, waiting to respond until a deepfake or a micro-targeted bot campaign has gone viral means you have already lost. The West is shifting toward predictive AI tools:
Tracking the Trajectory: Defense agencies utilize specialized software to monitor global data flows for "coordinated inauthentic behavior." Rather than just looking at the text of a post, these systems analyze the cross-platform speed, emotional framing, and structural amplification of a narrative to flag a Chinese cognitive operation while it is still in its infancy.
Repsense
Pre-bunking vs. Debunking: Psychological studies backed by NATO show that trying to correct a lie after the fact is highly ineffective. Instead, Western commands are experimenting with "pre-bunking"—alerting target populations or military units beforehand about the tactics and specific angles an adversary is likely to use to manipulate them, effectively immunizing the target cognitively.
