AI and the next revolution in military defence: China Daily columnist

A general view shows China's 19-year-old Go player Ke Jie (centre, left) competing during the second match against Google's artificial intelligence programme AlphaGo, on May 25, 2017.
A general view shows China's 19-year-old Go player Ke Jie (centre, left) competing during the second match against Google's artificial intelligence programme AlphaGo, on May 25, 2017. PHOTO: AFP

HUNAN (CHINA DAILY/ASIA NEWS NETWORK)- Last week, AlphaGo, an artificial intelligence (AI) programme developed by Google's DeepMind team, beat China's top go player Ke Jie in Wuzhen, East China's Zhejiang province, which suggests AI will occupy centre stage in the next technological and industrial revolution.

But what is AI's significance for the defence industry? And how to address the security challenges that AI may cause if inappropriately used in the defence sector?

AI is intelligence exhibited by machines that are made to think, work and react like humans in given situations. AI today is a combination of big data, cloud computing and internet of things.

Colossus was a set of computers developed by British codebreakers during World War II to help in the cryptanalysis of the Lorenz cipher. It is thus regarded as the world's first programmable, electronic, digital computer.

Since then, technologically advanced satellite sensors and wearable equipment have helped improve intelligent sensor and information processing. To manage data and information for maximum advantage, the development of command-and-control system has been on top of the military agenda.

AI has the potential to reshape defence technology and turn science fiction into reality. And the new challenges for defence and security should not be underestimated.

First, AI will make future wars more precise, quicker and crueller. The unmanned system will be widely used by the technologically advanced states to reduce their military casualties, which could pose a greater threat to civilians. And an AI war may undermine the rules of human morality.

Second, AI may never replace human ingenuity. So human ingenuity will be more important than the AI it gives birth to. Like any tool used in the wrong way, AI can cause a lot of harm and destruction.

Third, AI is far from being perfect. In a confrontational environment, AI systems could be invaded or damaged by malware and viruses. Thus, an uncontrollable AI system could end up killing innocent people.

And fourth, because of the coexistence of human-in-the-loop and human-out-of-loop, it is difficult to identify those responsible for faulty missions. For example, those responsible for a humanitarian catastrophe caused by drones may involve the pilot, the computer programmer, the procurement officer, the commander and more. Therefore, it would be more difficult to identify those responsible for AI mistakes than human errors.

Moreover, terrorists could exploit AI and its weaknesses to launch attacks. In an increasingly interconnected world, a cyber attack using AI technology could lead to the hacking of the control systems for nuclear weapons, or chemical and biological weapons.

Many countries are upgrading their defence capabilities by introducing AI technology, but they need to exercise extreme caution while using AI. And since it will become increasingly difficult for a country on its own to ensure absolute security, people should work more closely to respond to international security challenges. To build a community of shared destiny, I have three recommendations.

First, world leaders in military technology should take on more responsibility. A double-track security dialogue should be organised under the aegis of think tanks from all countries to discuss how to strengthen information exchange and sharing, assess the hacking risks in internet of things, estimate the effect of AI proliferation, and prevent wars and conflicts caused by both a sovereign state's miscalculation and attacks launched by non-state actors.

Second, in the disaster warning mechanism, global fight against terrorism, crackdown on transnational crime, and the humanitarian assistance missions, every country needs to facilitate deeper cooperation with other countries by using AI technology.

And third, drafting an international law on the use of AI as weapon is very important for avoiding AI abuse. Plus, international organisations should coordinate and cooperate to reach a consensus on how best to implement such a law globally.

The author is director and professor of the centre for National Security and Strategic Studies, National University of Defence Technology, China.