An article for PLA Daily said the technology has to be a ‘tool guided by human judgment’ and warns of its limitations. — SCMP
China’s military has warned that artificial intelligence cannot replace human decision-making on the battlefield because it lacks self-awareness.
“AI must work in tandem with human decision-makers to optimise command effectiveness, enhancing rather than replacing human agency,” an article published in People’s Liberation Army Daily on New Year’s Eve said.
The article says AI can be used to augment human capabilities – through data analysis, simulations or planning – but cannot replace them.
“As AI evolves, it must remain a tool guided by human judgment, ensuring accountability, creativity, and strategic adaptability remain at the forefront of military decision-making,” the article said.
The article said human autonomy and creativity are indispensable on the battlefield. It added that while human commanders can respond to situations dynamically, and exploit enemy weaknesses, AI operates within predefined algorithmic boundaries and its responses often lack originality.
The PLA said it preferred a model where “humans plan and AI executes” in which the technology is used to carry out the strategies and tactics developed by commanders and keeps an element of human oversight.
It said it prefers a structure where machines analyse data, provide insights and suggest potential actions. However, “final decisions rest with human commanders, safeguarding against errors stemming from AI’s black-box nature”.
The article said another weakness of the technology is its inability to reflect on its own actions or take responsibility for its decisions, unlike human commanders who can refine their plans in response to the circumstances.
The article said even advanced systems, such as autonomous missile defence, typically rely on human operators for final decisions to ensure accountability.
Beijing is a strong advocate of regulating the military use of AI and has previously submitted papers to the United Nations on this theme as well as stressing the need for “human control over the technology”.
President Xi Jinping has also stressed the need to control the technology, and in his most recent meeting with his US counterpart Joe Biden in Peru two months ago the two agreed that “human beings and not artificial intelligence should make decisions over the use of nuclear weapons”.
The Pentagon’s latest annual report into the Chinese military, published last month, said the PLA was working on human-machine collaboration and was aiming for “algorithmic” and “network-centric” warfare by 2030.
The report noted the PLA’s significant investments in autonomous vehicles, predictive maintenance, automated target recognition and submarine drones. It also said it was planning to use AI-boosted sensors to improve missile accuracy and exploring ways to use the technology for war games and training.
The report also said the PLA was looking to use civilian AI innovations in areas such as drones, image recognition and intelligent decision-making.
The United States military is looking at ways it can use the technology. Last month the Pentagon’s chief digital and AI officer Radha Plumb unveiled details of a scheme to integrate the technology into combat operations, which she described as the “first major effort to deploy leading edge AI to support warfighter needs in real time”.
She also warned that the US of the technology by “adversaries” such as China, Russia, Iran and North Korea “is accelerating and poses a significant national security risk” to the US. – South China Morning Post
