New English Edition
Artificial Intelligence in Horticulture
Kim Kyung-jin, Attorney at Law
Across five chapters and ten sections, this book examines computer vision for crop diagnosis, harvesting robots and autonomous field systems, smart greenhouses and digital twins, precision irrigation and supply-chain quality control, high-throughput phenotyping, and predictive breeding.
New English Edition
Artificial Intelligence in Food Crop Agriculture
Kim Kyung-jin, Attorney at Law
Across six chapters and eighteen sections, the book examines digital agricultural infrastructure, remote sensing, crop diagnosis, yield forecasting, precision irrigation, genomics, molecular breeding, agricultural robotics, climate-smart agriculture, and global food security.
New English Edition
The Future of Forestry and Agroforestry
Kim Kyung-jin, Attorney at Law
Driven by Artificial Intelligence and Digital Innovation
Across five chapters and fifteen sections, the book follows satellites, drones, LiDAR, digital twins, forest-specific language models, wildfire and pest forecasting, forestry robotics, agroforestry, timber traceability, and forest carbon markets.
New English Edition
Smart Livestock Farming: AI Enters the Barn
Kim Kyung-jin, Attorney at Law
Sensors listen, cameras watch, and artificial intelligence helps farmers decide.
Across five chapters and fifteen sections, the book follows precision livestock farming from animal health and reproduction to robotic milking, virtual fencing, digital twins, methane reduction, welfare, and data ownership.
Table of Contents
Han Dong-hoon, Busan Buk-gu Gap: A Record of the 100 Days Before and After the Election (Mar. 26-Jul. 3, 2026)
Kim Kyung-jin
Table of Contents and 13 sections
From March 26 to July 3, 2026, this record follows the spring after expulsion, the Busan Buk-gu Gap by-election, victory as an independent, and the first bill submitted in the National Assembly.
Table of Contents
Artificial Intelligence and Medicine
Kim Kyung-jin, Attorney at Law
AI in clinical care, hospitals, education, and research
AI in medical imaging, risk prediction, treatment planning, hospital operations, education, and research, with patient safety, privacy, and accountability.
[AI Library] Chapter 34. Hacking and Deception
Chapter 34. Hacking and Deception
Hacking and Deception I remember one night when I was flying in pitch darkness at an altitude of 30,000 feet. The RWR, or Radar Warning Receiver, suddenly started screaming. The enemy's surface-to-air missile radar caught me. There were only two things I could believe in at that moment. The numbers on the dashboard flashed before my eyes, and it was my intuition accumulated from thousands of hours of flight experience. When the numbers lie, your intuition will tell you. When your intuition wavers, numbers set you straight. The balance between these two allowed me to land alive. But AI pilots have no intuition. This is the crux of the problem.
AI sees the world through data. Signals sent by radar, heat sources picked up by infrared sensors, and friendly forces' location information received through data links. All of this is input into the AI’s brain as a combination of 0 and 1. If the input value is accurate, AI makes decisions faster and more accurately than humans. But what if the input value is false? AI accepts the lie as truth and runs in the wrong direction at a much faster speed than humans. In ancient Greek mythology, there is a hero named Achilles. He was an invincible warrior. Because his mother made him immortal by dipping him in the River Styx.
But only the heel with which his mother was holding him did not touch the water. In the end, Achilles was hit by an arrow in his heel and fell. The lesson is that even the strongest have weaknesses. The heel of AI pilot is ‘data’. If an enemy can manipulate or corrupt this data, the invincible AI warrior will be powerless in an instant. Let me start with the most horrifying threats: hostile attacks, hallucinations with your eyes open. Experts call this an ‘Adversarial Attack’. It may sound difficult, but let me explain it to you. When AI recognizes objects, it analyzes the patterns of pixels, or tiny dots that make up the screen.
By looking at millions of photos, it learns, “If this pattern is a tank,” or “If that pattern is a truck,” it learns. But what happens if an enemy draws very fine patterns on the surface of the tank that are barely noticeable to the human eye? The pattern is a ‘customized poison’ created by reverse analysis of AI’s neural network structure. To the human eye, it still appears to be a tank. However, in the eyes of AI, the tank is recognized as a ‘school bus’, a ‘flock of sheep’, or simply a ‘rock’.
Imagine this. I give orders to my AI wingman. “Attack enemy tanks in front.” But the AI answers. “There are only civilian vehicles ahead.” I'm clearly looking at a 60-ton steel monster, but my AI colleague next to me doesn't see it. This is not a malfunction of the machine. The machine is operating normally. However, the enemy has created an illusion in the machine's eyes. A similar thing actually happened at the Chinese People's Liberation Army's training ground in 2024. It was believed that the Blue Team's AI-assisted fire units had accurately hit the Red Team's batteries.
But when the training controller stopped the situation, the truth that emerged was devastating. More than half of BLU's fire force had already been destroyed. The red team commander had placed fake batteries and fake signals at the shooting range. The Blue Team's AI revealed its location by shooting at the ghost. The opposite situation is more dire. What if an enemy projects a special pattern on the roof of a private hospital or school, causing the AI to mistake it for a ‘missile launch pad’? Our AI wingman will hit there without hesitation. It's a mistake. Civilians die. And who is responsible for that? This is not just a technical issue.
It is a problem that shakes the very nature of war. Data contamination, poisoning the well There are more insidious attacks. This is an attack that takes place before the war even begins. Experts call this 'Data Poisoning'. It's the same way as secretly poisoning a well. AI grows by eating data. It learns by analyzing millions of simulations, tens of millions of photos, and hundreds of millions of sensor data. But what would happen if an enemy cyber unit secretly infiltrated our training data server and planted a very subtle error?
For example, an enemy could 'train' an AI so that it behaves normally in normal times and then malfunctions only under certain conditions. It only behaves strangely when it detects a red smoke bomb, when it hears radio waves of a certain frequency, or only on a certain date. This is called a ‘backdoor’. AI, which normally operates without any problems, suddenly becomes out of control at a critical moment when war breaks out. They mistake their allies for enemies, ignore enemy planes and pass by, or crash completely. This is a digital version of a Trojan horse.
The ancient Greek allies hid their soldiers inside a giant wooden horse and brought them into the city of Troy. The Trojans saw the wooden horse as a trophy and willingly
I accepted it. At night, soldiers came out of the wooden horse and opened the city gates, and Troy was destroyed. If the enemy has hidden a 'horse' in our AI's learning process, we will deploy that AI into the battlefield without knowing it. And at a critical moment, the poison in the wooden horse kicks in. Training military AI requires massive amounts of data, including satellite imagery, reconnaissance images, and open source information. It is nearly impossible to completely verify the purity of all this data. Enemies can distribute large quantities of fake images of military equipment on the Internet.
When our AI learns those images, it will not be able to properly recognize the enemy's real weapon. A strategy that turns our most powerful weapons into scrap metal without ever firing a shot. Spoofing, Whispering a False Reality In the past, electronic warfare was about blinding the enemy's ears with loud noise. The powerful radio interference disrupted the radar screen like a blizzard. But electronic warfare in the AI era is different. It is a hypnotism that whispers ‘false reality’ to the enemy. Let's take GPS spoofing as an example. GPS calculates your location by receiving signals from satellites.
But what if the enemy sends a fake GPS signal that is stronger than the satellite? AI pilots think they are in a completely different location than they actually are. It flies believing it is 3 kilometers away. They drop bombs thinking they have reached the target, but in reality it is an empty field with nothing in it. Or even worse, a friendly base. GPS spoofing has already become commonplace on the Ukrainian battlefield. Drones lose their location and fly to strange places or crash. As of 2025, the Russian military is conducting a large-scale GPS jamming operation over Ukraine. Hundreds of drones are affected.
Data link hacking is even scarier. The AI pilot is connected to the friendly command center via a data link. Mission orders, target coordinates, and friendly location information are transmitted through this link. What if an enemy hacks this link and sends fake commands? “Formation, immediately turn 180 degrees and return.” This command came through an encrypted channel. The AI turns its nose without question. A human pilot would probably wonder, "It's strange to be ordered to return in this situation." We will request re-confirmation via voice command. However, AI will carry out commands as long as there is no logical contradiction.
Inflexibility is the limitation of machines.
Hacking takes over the digital cockpit We are now sitting in a ‘digital cockpit’. Modern fighter jets like the F-35 and F-47 are giant flying computers. Millions of lines of code make the aircraft move. What if an enemy hacker comes through the network and takes control of the AI system? It's like the enemy is riding in my back seat and holding a knife to my throat. There is no such thing as perfect code in the software world. This is why your computer or smartphone is updated frequently. This is to belatedly close security holes that developers did not discover. The same goes for combat prayers.
No matter how thoroughly you inspect, there may be gaps somewhere. If the enemy finds the gap first, the F-47 will become a flying $100 billion brick or a zombie drone controlled by the enemy. The problem is more serious in the case of drone swarms, or swarms of drones. Hundreds of drones are networked together and move like one giant creature. But that link becomes a weakness. Even just one hack can compromise the entire colony. Just as a virus spreads throughout the body, malicious code spreads across networks. The U.S. Defense Advanced Research Projects Agency (DARPA) is taking this issue seriously.
In 2025, they launched a project called SABER. It is an abbreviation for ‘Securing Artificial Intelligence for Battlefield Effective Robustness.’ The goal of this project is to assess the vulnerability of AI systems against adversarial attacks, cyberattacks, and electronic warfare attacks and develop defense technologies. DARPA officials warned that there is currently no ecosystem that can systematically assess the security vulnerabilities of AI systems deployed on the battlefield. The breakdown of trust is what I fear most.
As a fighter pilot, my greatest fear is not the enemy. ‘It’s a situation where I don’t trust my airplane.’ The moment you suspect that your AI wingman has been hacked or deceived, he or she is no longer a force but a potential threat. When I flew the Wild Weasel mission over Iraq, I trusted both my dashboard and my eyes. If the dashboard showed strange numbers, I checked it with my eyes. If I saw something strange, I checked the dashboard again. These two validated each other. But the AI pilot only knows what the sensors show. Even if the sensor is telling a lie, you can't 'feel' it.
To win the algorithm war, we must secure not only AI's performance but also its 'robustness'. We need AI that is not fooled by the enemy's deception tactics, is unshaken by data contamination, and can defend itself against hacking attempts. The US Department of Defense operates the ‘AI Red Team’. They are a group of experts who ruthlessly attack and hack friendly AI systems to find vulnerabilities. We make AI endure tens of thousands of digital tests before it is deployed in the field. In the end, no matter how advanced technology is, humans must be the last resort.
When the AI says, “This is a truck,” the pilot must be able to use his intuition and experience to determine, “No, that’s a camouflaged tank.” Amid threats of hacking and deception, human intuition is the last safety device that catches AI errors. This is why we cannot completely rely on machines. To survive on the battlefield, you must be able to doubt what you see. AI hasn't learned that suspicion yet. That is the most critical weakness of algorithmic warfare. The ruler of the skies is no longer the fastest. You will be the one who is least deceived.
Kim Kyung-jin
Attorney · Former Member of the National Assembly · AI Policy Researcher
© 2026 Kim Kyung-jin. All rights reserved.