AI and the Future of Warfare

#1 · 17th October 2021, 11:38 pm

Quote from Shayan Hassan Jamy on 17th October 2021, 11:38 pm
Introduction. Often, when people think about the use of Artificial Intelligence (AI) in warfare, their mind goes to the distant future. AI and warfare, however, are fast becoming inseparable terms. This was manifested on November 27^th, 2021, when Iranian nuclear scientist Mohsen Fakhrizadeh was assassinated by Israeli forces. Although initial claims about the operation were varied and often conflicting, a recent report by the New York Times (NYT) states that the assassination was carried out by an ‘AI-assisted remote control killing machine’. Although the notion of AI having been used in this case had been speculated upon earlier, it was unclear exactly what role it had played in the killing. The NYT report, however, gave a detailed view of the entire operation in which both human operators and AI played a significant role.
The Killing of Iranian Scientist.          Dr. Fakhrizadeh had been considered by Israel’s Mossad as the ‘father of the Iranian nuclear program’ and had been identified as an important target for many years. Since 2007, five scientists connected with Iran’s nuclear program had been assassinated by Israeli forces. Dr. Fakhrizadeh had remained elusive. On the day of the operation, he decided to ignore his bodyguards’ orders and drive his car himself. Along his route, he was shot and killed by a remote-controlled machine gun mounted on a pick-up truck. His wife, who was in the passenger seat beside him, remained unharmed. When his bodyguards arrived on the scene, there were no assailants to be found. The machine’s operators had, in fact, carried out the attack from more than 1000 miles away.
The machine’s various parts, the report claims, were brought into Iran separately and assembled there. It was then fixed on a pick-up truck which also had multiple cameras attached to it. A group of operators used the cameras and a satellite link to control the machine from a distance. Even with this technology, there was still a slight delay between the operator and the machine, which is where AI came into play. According to the NYT report, AI was used to compensate for a delay of 1.6 seconds between the camera images reaching the machine operator and the operator’s response reaching the machine. AI was also used to adjust the machine to allow for the truck shaking from the machine gun’s recoil, and for the speed of Dr. Fakhrizadeh’s car.
Once Dr. Fakhrizadeh was positively identified by the operators through the camera images, the gun was fired upon him multiple times. When the Iranian bodyguards arrived at the scene, they could only find half-destroyed machine parts. Explosives which were meant to destroy it and any evidence were the only part of the operation which failed. Because of the fact that Dr. Fakhrizadeh’s wife was unharmed, Iranian investigators initially believed that facial recognition software had also been used. However, this was not confirmed in the NYT report.
Future of AI in warfare.         The use of AI in this assassination might not seem that alarming to many people, but it deserves a closer look. AI was used here to minimize human errors and act only as a supplementary to human operatives. While most people worry about fully-autonomous weapons when it comes to AI, the scope of AI in warfare is broader than that. Experts predict that AI will become integrated into various aspects of defence within the next few decades, from actual warfare to cybersecurity and decision-making. Already, AI is blurring the lines of warfare. Technology that it being developed purely for civilian purposes, such as facial recognition technology, could easily be repurposed and used to engage the enemy. As we have seen historically, states that take a lead over others in understanding and implementing a certain technology militarily often become major global players. The British, after the Industrial Revolution or the US after acquiring nuclear weapons serve as good examples. Whoever takes the lead in AI could very well become the dominant superpower for years to come.
The normalization of AI in warfare is already taking place. Powers such as the US, China, and Russia, have already incorporated AI into their defence, and this trend does not show any signs of stopping. A report released by the US National Security Commission on Artificial Intelligence in March 2021 made several AI-related recommendations. This included advising that the Pentagon increase its AI defence spending from $1.5 billion to $8 billion by 2025, and for the US military to implement AI throughout its ranks. This was seen as a direct response to the AI advancements of other countries like China. China, itself, has stated that it wants to be the ‘global AI leader’ by 2030 and is in many ways ahead of the US in terms of AI research. Russia has also stated certain AI-related goals that it wishes to achieve by 2030. India, Israel, the UK and others have also taken steps indicating they will invest heavily on AI in the future.
International discussion regarding the use of AI in warfare is, unfortunately, very limited. Major states disagree on most aspects and the current rate of technological advancement far outweighs the speed of international diplomacy to address its potential problems. This could have a number of far-reaching consequences. Firstly, with the rapid advancements being made in AI, new threats could evolve regularly. This could encourage states to spend more on AI and ultimately lead to an arms race. Also, the possible large-scale proliferation of AI-enabled technologies would be much more difficult to control than in the past. Since the major component of AI is the algorithm, which is intangible software, it is much easier to obtain for both state and non-state actors. There is a host of other problems that could arise in the future but which we cannot even formulate at this time. Lethal autonomous weapons systems are not as far away as many believe.
What is required is a robust international body that can address major AI concerns as and when they arise. Currently, however, there is a lack of both consensus and will on the part of many major states to take the initiative in this regard. This is worrying, and more needs to be done to combat the potential problems that the use of AI in warfare could bring.

Introduction. Often, when people think about the use of Artificial Intelligence (AI) in warfare, their mind goes to the distant future. AI and warfare, however, are fast becoming inseparable terms. This was manifested on November 27^th, 2021, when Iranian nuclear scientist Mohsen Fakhrizadeh was assassinated by Israeli forces. Although initial claims about the operation were varied and often conflicting, a recent report by the New York Times (NYT) states that the assassination was carried out by an ‘AI-assisted remote control killing machine’. Although the notion of AI having been used in this case had been speculated upon earlier, it was unclear exactly what role it had played in the killing. The NYT report, however, gave a detailed view of the entire operation in which both human operators and AI played a significant role.

The Killing of Iranian Scientist. Dr. Fakhrizadeh had been considered by Israel’s Mossad as the ‘father of the Iranian nuclear program’ and had been identified as an important target for many years. Since 2007, five scientists connected with Iran’s nuclear program had been assassinated by Israeli forces. Dr. Fakhrizadeh had remained elusive. On the day of the operation, he decided to ignore his bodyguards’ orders and drive his car himself. Along his route, he was shot and killed by a remote-controlled machine gun mounted on a pick-up truck. His wife, who was in the passenger seat beside him, remained unharmed. When his bodyguards arrived on the scene, there were no assailants to be found. The machine’s operators had, in fact, carried out the attack from more than 1000 miles away.

The machine’s various parts, the report claims, were brought into Iran separately and assembled there. It was then fixed on a pick-up truck which also had multiple cameras attached to it. A group of operators used the cameras and a satellite link to control the machine from a distance. Even with this technology, there was still a slight delay between the operator and the machine, which is where AI came into play. According to the NYT report, AI was used to compensate for a delay of 1.6 seconds between the camera images reaching the machine operator and the operator’s response reaching the machine. AI was also used to adjust the machine to allow for the truck shaking from the machine gun’s recoil, and for the speed of Dr. Fakhrizadeh’s car.

Once Dr. Fakhrizadeh was positively identified by the operators through the camera images, the gun was fired upon him multiple times. When the Iranian bodyguards arrived at the scene, they could only find half-destroyed machine parts. Explosives which were meant to destroy it and any evidence were the only part of the operation which failed. Because of the fact that Dr. Fakhrizadeh’s wife was unharmed, Iranian investigators initially believed that facial recognition software had also been used. However, this was not confirmed in the NYT report.

Future of AI in warfare. The use of AI in this assassination might not seem that alarming to many people, but it deserves a closer look. AI was used here to minimize human errors and act only as a supplementary to human operatives. While most people worry about fully-autonomous weapons when it comes to AI, the scope of AI in warfare is broader than that. Experts predict that AI will become integrated into various aspects of defence within the next few decades, from actual warfare to cybersecurity and decision-making. Already, AI is blurring the lines of warfare. Technology that it being developed purely for civilian purposes, such as facial recognition technology, could easily be repurposed and used to engage the enemy. As we have seen historically, states that take a lead over others in understanding and implementing a certain technology militarily often become major global players. The British, after the Industrial Revolution or the US after acquiring nuclear weapons serve as good examples. Whoever takes the lead in AI could very well become the dominant superpower for years to come.

The normalization of AI in warfare is already taking place. Powers such as the US, China, and Russia, have already incorporated AI into their defence, and this trend does not show any signs of stopping. A report released by the US National Security Commission on Artificial Intelligence in March 2021 made several AI-related recommendations. This included advising that the Pentagon increase its AI defence spending from $1.5 billion to $8 billion by 2025, and for the US military to implement AI throughout its ranks. This was seen as a direct response to the AI advancements of other countries like China. China, itself, has stated that it wants to be the ‘global AI leader’ by 2030 and is in many ways ahead of the US in terms of AI research. Russia has also stated certain AI-related goals that it wishes to achieve by 2030. India, Israel, the UK and others have also taken steps indicating they will invest heavily on AI in the future.

International discussion regarding the use of AI in warfare is, unfortunately, very limited. Major states disagree on most aspects and the current rate of technological advancement far outweighs the speed of international diplomacy to address its potential problems. This could have a number of far-reaching consequences. Firstly, with the rapid advancements being made in AI, new threats could evolve regularly. This could encourage states to spend more on AI and ultimately lead to an arms race. Also, the possible large-scale proliferation of AI-enabled technologies would be much more difficult to control than in the past. Since the major component of AI is the algorithm, which is intangible software, it is much easier to obtain for both state and non-state actors. There is a host of other problems that could arise in the future but which we cannot even formulate at this time. Lethal autonomous weapons systems are not as far away as many believe.

What is required is a robust international body that can address major AI concerns as and when they arise. Currently, however, there is a lack of both consensus and will on the part of many major states to take the initiative in this regard. This is worrying, and more needs to be done to combat the potential problems that the use of AI in warfare could bring.