Chinese PLA soldiers practice with FPV drones
China’s new DeepSeek chatbot is shaking things up in the U.S. stock market. It’s said that creating the DeepSeek-V3 model cost around $6 million, which is just a fraction of what similar U.S. systems require, using far less computing power.
This release came right after the U.S. announced big plans to invest in AI infrastructure. DeepSeek’s creators argue that such a massive investment is now unnecessary. The launch has been labeled a “Sputnik moment,” similar to when the Soviet Union surprised the world by launching the first satellite in 1957, putting them ahead in the space race.
While Sputnik’s launch had clear military concerns, China’s edge in advanced AI could potentially shift the balance in warfare. A new generation of AI-driven drones could be developed, giving China a significant advantage.
Drone Dominance With AI
DJI Sky City, the global headquarters of DJI, a major player in the consumer drone market.
China stands as the leading manufacturer of small drones, with DJI controlling about 80% of the global consumer drone market. Despite DJI’s stance against military uses of their products, their drones have become essential for both sides in the Ukraine conflict.
Affordable FPV drones, often equipped for combat, have emerged as vital tools in the ongoing war. While these drones are assembled locally, many parts originate from China.
Currently, most drones are manually operated by skilled pilots. For instance, Tymofiy Orel, a 25-year-old pilot, earned the title “Hero of Ukraine” after taking out a significant number of enemy tanks and vehicles.
In recent times, small drones have started to incorporate more AI features. American firm Auterion is providing Ukraine with compact computers, like the Skynode-S boards, designed for machine learning. This tech could help future pilots achieve success similar to Orel’s.
Former Ukrainian Defense Chief Valerii Zaluzhnyi recently acknowledged the rapid technological advancements and how AI has begun to play a crucial role in modern warfare.
“When robots first appeared on the battlefield in 2022, many dismissed them. Yet by 2024, AI’s impact on warfare became undeniable,” he stated.
In the coming year, AI-enabled drones are expected to rise on the battlefield, but challenges remain, such as equipping drones with sufficient computing power and training AI systems to function effectively.
Transforming Data into Action
LLMs like DeepSeek are finding new uses in drone warfare.
DeepSeek utilizes a Large Language Model (LLM), a type of AI trained on extensive text data that’s great for natural language interactions. Surprisingly, LLMs can also improve drone functions. Research, primarily from China, has explored how LLMs can make drones smarter and more efficient.
A 2024 study showcases how integrating LLMs can help drones process data and make quick decisions. While specifics are avoided, the authors note that LLMs excel at recognizing objects, individuals, and activities in visual data—crucial for military and civilian use.
For example, drones can execute complex tasks like checking for vehicles along a specific route or approaching an enemy location from a hidden angle.
Not only can drones complete orders, but they can also report findings back in natural language, enhancing communication. Imagine a drone saying, “No vehicles detected, one destroyed truck,” or “Enemy positions show no activity, but heat signatures indicate they are occupied.”
LLMs also allow drones to communicate with each other, enabling teamwork without needing humans to direct them. A reconnaissance drone might relay tasks to attack drones while continuously updating them on mission progress.
Another Chinese research paper outlines how LLMs could lead to fully autonomous combat drones. They envision a scenario where a single human pilot collaborates with robotic drones that react swiftly and communicate as humans do.
DeepSeek’s Potential
DeepSeek is touted to be more efficient than other LLMs, needing less computing power—ideal for smaller applications like drones. Its developers also claim it offers fast results with minimal training data.
While the true capabilities of DeepSeek remain uncertain, a cost-effective AI system for drones could be a significant asset for China right now.
China’s stronghold on drone manufacturing, combined with insights gained from the Ukraine conflict, positions them favorably. They are actively building their own combat drone units, often using affordable commercial drones.
With LLM technology in the mix, China could dominate drone warfare. As the landscape shifts towards drone vs. drone confrontations, the side with smarter software may determine the victor.
Source link
LLM,Deepseek,drone,FPV,smart drone