To ban or not to ban

Russian aggression against Ukraine which began on February 24 makes the question of how Poland should get prepared for a potential conflict more and more important. The significance of new weapons like unmanned aircraft or cybersecurity attacks is already evident in this war. The sheer number of troops does not guarantee a quick victory. The armored conflicts of the future will be even more different from the wars fought between armies of hundreds of thousands of soldiers in the 19th or 20th century. Most likely they will be fought not by numerous but highly skilled professionals supported by cutting-edge technology, most likely including autonomous, AI-controlled weapons.

Many people argue that, for ethical reasons and because of the danger of uncontrolled proliferation, autonomous lethal weapons should be fully banned, similar to chemical or biological weapons. The topic of banning lethal autonomous weapon systems (AWS) has been coming back many times during the last decade. The recent talks at the United Nations to regulate AI-controlled weapons failed again, as some of the countries leading in the military development, including the United States and Russia, were against such regulations.

Back in 2013, activists and NGOs from many countries launched the “Campaign to Stop Killer Robots.” The initiative was supported by 3,000 robotics and AI experts who issued an open letter calling for a complete ban on offensive AWS operating without human control.

Such autonomous weapons already exist, for example, Israeli Harpy or Turkish Kargu drones. They can be launched to fly over a particular area, search for a target and destroy it when found. Unlike nuclear weapons, AI-controlled weapons are very cheap and easy to construct. In the future, the miniaturized, fly-size drones equipped with face recognition systems may become cheap and practically unstoppable assassins.

But there are also benefits of AI-controlled weapons. The supporters of AI weapon systems argue that their use will reduce the number of casualties, first soldiers, but also civilians. AI systems do not have the mental or physiological limitations that humans have, they are not subject to stress and emotions such as fear or hysteria leading to irrational behavior. The rape of women by soldiers that occurred throughout the history of wars will be a thing of the past.

Autonomous weapons have also clear technical advantages. Currently, constructed airplanes have numerous limitations, which are imposed only to protect the human body. A human can bear only certain accelerations and overloads related to them. After eliminating human pilots, these constraints would not be needed anymore. The plane would be able to withstand greater overloads, which would allow increasing its maneuverability. And of course, the pilot's life wouldn’t be risked.

For many smaller nations, using AI-controlled weapons may be a way to reduce the military supremacy of their more powerful hostile neighbors. In Poland, AWS, although not widely present in a political debate, is still a subject of discussions conducted by geopolitical thinktanks.

Regardless of ethical issues, the leaders in AI — China and the US — are determined to keep developing military applications of AI systems. While the arms race of the 20th century was dominated by the development of weapons of increasing power, now the arms race is shifting towards AI-controlled weapons in all areas of the potential battlefield, including outer space and cyberspace.

Although it may be extremely difficult to regulate AI-controlled weapons, it is still worth trying. The armed conflicts of the future will look completely different than today. They may be more deadly than nowadays or just the opposite, depending on how the autonomous weapons will be handled.

Jacek Krawczak is a senior data scientist at Atos R&D Poland and an AI senior expert at Atos Global Expert Community. Founded in 1997, Atos is a French multinational IT service and consulting company.