discussion | how far away is the humanoid robot from the battlefield?
2024-09-27
한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina
■zhang yicheng
not long ago, a startup invested by openai released a humanoid robot designed for home environments, and soon after launched a virtual world model for developing and training robots. in the introduction video, the robot's movements are natural and smooth, and its movement speed is close to that of humans. many netizens once thought that it was a real person pretending to be a real person. some industry insiders claim that humanoid robots are about to have their “chatgpt moment.”
engels pointed out: "humans fight in the same way they produce." in some local conflicts in recent years, the role of intelligent weapons and equipment such as drones and robot dogs has become increasingly prominent. so, as the best carrier of artificial intelligence in the physical world, how far away are humanoid robots from the battlefield?
first of all, it should be noted that the battlefield application of humanoid robots has become a general trend. the human form is the only specific form of advanced intelligent creatures produced by natural evolution, and has withstood the harsh test of the environment over a long period of time. designing robots into humanoid shapes can better adapt to human working environments and weapons and equipment, perform diverse military tasks, have greater versatility, and achieve higher battlefield utilization. in the past, limited by technology, the battlefield application of humanoid robots was once limited. however, with the rapid advancement of large-scale model technology, the new generation of humanoid robots already have a "brain" that can perceive complex scenarios such as decision-making and planning, and the ability to accurately control high-degree-of-freedom body movements. "cerebellar", the ability to adapt to complex battlefields has been significantly improved. in addition to the strong traction of battlefield needs and the acceleration of technological breakthroughs, the explosive development of the industry also provides solid support for the application of humanoid robots on the battlefield. the current market demand for humanoid robots is growing rapidly, and it is estimated that by 2035, the global market size is expected to reach us$154 billion. a u.s. company plans to start small-batch production of humanoid robots in 2025, with a long-term goal of annual production of 100 million units.
it is foreseeable that humanoid robots will have increasingly diverse battlefield application scenarios. after-loading support may be one of the first application scenarios for humanoid robots. relatively speaking, in application scenarios such as warehouse management, material handling, equipment maintenance, weapons manufacturing, and dangerous goods destruction, the tasks that need to be completed by humanoid robots are simple and repetitive, and the environment must be closed. humanoid robots at this stage are already competent and can significantly improve work efficiency and reduce safety hazards. the second is assisted decision-making. artificial intelligence-assisted combat decision-making is a significant symbol of intelligent warfare. humanoid robots are ideal mobile artificial intelligence decision-making assistance systems because of their powerful computing power, flexible maneuverability and friendly human-computer interaction. they can provide intellectual support for human decision-making in combat and ensure a decision-making advantage against the enemy. the highest level application scenario for humanoid robots is collaborative operations. humanoid robots can not only fight side by side with humans and assist in performing various tasks, but can also act as decoys to attract firepower when necessary, actively "sacrifice" to cover humans. this requires humanoid robots to not only have excellent military skills such as mobile concealment, intelligence reconnaissance, identification and positioning, and weapon control, but also have a variety of human-computer interaction methods such as language, movement, and emotion. the most important thing is that they must have safe and reliable ability to make independent decisions.
however, the current battlefield application of humanoid robots still faces many challenges. first, key technologies are in urgent need of breakthroughs. for example, in terms of large model empowerment, there are problems such as the lack of high-quality battlefield data for training models and the relatively slow movement of control robots. in order to increase the use time, breakthroughs in high-energy-density batteries and ultra-lightweight materials are needed. secondly, moral ethics are challenged. un secretary-general antonio guterres has said that machines with the ability and discretion to take human lives are "politically unacceptable, morally repugnant and should be banned under international law." accidental bombing accidents caused by humanoid robots will also lead to liability dilemmas. long-term interaction between humans and humanoid robots will lead to cognitive and psychological deviations. the third is that costs need to be significantly reduced. currently, the mainstream commercial humanoid robots on the market cost millions of yuan per unit, and the cost for military use is even higher. if we want to achieve large-scale application on the battlefield, we must significantly reduce production costs by stabilizing the manufacturing process, reducing parts costs, and commercializing mass production.
(source: china military network - people's liberation army daily)
