“Compared with previous AI models, ChatGPT is more human-like and has many potential applications in the military field,” Mao Weihao, a lecturer with the Army Command College in Nanjing, wrote in the article published on the newspaper’s website on Thursday.

He said the software could be used as a battlefield intelligence system and to plan cognitive warfare.

During peacetime, ChatGPT could be used to gather intelligence by helping to analyse vast amounts of data from the internet, improving efficiency and finding valuable information, he argued. In wartime, it could generate comprehensive battlefield reports, making planning more efficient.

Researcher blames China’s lack of a ChatGPT alternative on lack of vision

ChatGPT could also be used for cognitive warfare by analysing public opinion, creating false statements, and manipulating perceptions to damage a government’s image or change the public’s stance, the article continued.

Mao noted that AI had already been used in some military applications, such as radar systems and drone swarms.

“Armed forces of some developed countries utilise AI in radar systems for rapid tagging and individual recognition of airborne targets. Some have already achieved automatic detection and identification of small targets like drones,” Mao said.

“The US Department of Defence’s Defence Advanced Research Projects Agency is implementing a series of plans to research the use of drone swarms, autonomous underwater vehicles and ground-based mobile platforms to achieve respective combat objectives.”

Mao also suggested countermeasures against AI in war, saying that the technology had known vulnerabilities.

He said its performance depended on training data, but this could be targeted to deceive the AI. One example he cited was research from KU Leuven University in Belgium, which found that the software could not recognise someone as human if they were wearing a particular A4-sized colour pattern.

He also said AI could not create new knowledge and struggled to break convention or create new strategies, giving human creativity an advantage over the technology.

“The ability to think, to break conventions, and innovate tactics will be crucial for future battles against ‘AI commanders’,” he said.

As a way of testing ChatGPT , the Post asked its latest model, GPT-4, whether it had military uses.

Several Chinese experts back call by Musk and others for pause in AI race

It replied: “As an advanced AI language model, I have many potential applications in the military field”, listing areas such as intelligence processing, decision-making, training simulation, unmanned combat system operating, equipment maintenance and supply chain optimisation.

However, it also highlighted ethical and moral concerns, adding: “It is crucial for AI developers, governments, and international organisations to work together to establish responsible and transparent guidelines for AI deployment in military settings.”