Academia.eduAcademia.edu

Figure 10 – uploaded by Rishi Shah

See full PDFdownloadDownload figure
The use of the term AI therefore would represent systems made up of all of the above sub- systems. The GGE should consider artificially intelligent systems to be systems that take in percepts, make decisions based on those percepts and previous knowledge, and act accordingly to accomplish a certain goal. The discussion should still differentiate between narrow AI and general AI, distinguishing the ability of a system to perform a single or narrow set of tasks, as opposed to fully thinking and behaving like a human, the premise of general AI.

Figure 10 The use of the term AI therefore would represent systems made up of all of the above sub- systems. The GGE should consider artificially intelligent systems to be systems that take in percepts, make decisions based on those percepts and previous knowledge, and act accordingly to accomplish a certain goal. The discussion should still differentiate between narrow AI and general AI, distinguishing the ability of a system to perform a single or narrow set of tasks, as opposed to fully thinking and behaving like a human, the premise of general AI.

Related Figures (10)

The word autonomy comes from the early 17" century, Greek word, autonomia, stemming from autonomos which means, ‘having its own laws’. The Oxford English Dictionary defines autonomous as “freedom from external control or influence; independence.” Based on the above definition, LAWS are weapons systems that apply lethal force independently, with freedom fror human control. This definition encompasses weapon systems from landmines! to unmanned aeria vehicles” to the Israeli Harpy*. While this range of weapons systems have autonomous aspects they differ in the complexity of their algorithms and software, or their level of intelligence Landmines merely sense pressure, movement, sound, magnetism or vibration while the Harpy “detects, attacks and destroys enemy radar emitters, hitting them with high hit accuracy” [2]. The increasing complexity of weapons systems and their software led to the introduction of AI into the LAWS discussion. International discussions about lethal autonomous weapons systems (LAWS) have become more prevalent as nations attempt to come to terms with how these weapons work and how tc regulate them. The terms in which the community discusses them can at times be ambiguous as there is an absence of a common vocabulary and common understanding of the terms autonomy and artificial intelligence (AI). A weapon system is comprised of four actions, searching for < target, detecting that target, making the decision to engage the target, and engaging the target. [1 Developed by Colonel John Boyd, the OODA loop is a model for competitive decision making comprised of those four actions that define a weapon system: observe, orient, decide, and act. [1 In human operated weapons systems, a human operator completes all of these actions. In a weapor system with autonomy, the human operator relinquishes control over some part of the loop. autonomous as “freedom from external control or influence; independence.” Based on the above
The word autonomy comes from the early 17" century, Greek word, autonomia, stemming from autonomos which means, ‘having its own laws’. The Oxford English Dictionary defines autonomous as “freedom from external control or influence; independence.” Based on the above definition, LAWS are weapons systems that apply lethal force independently, with freedom fror human control. This definition encompasses weapon systems from landmines! to unmanned aeria vehicles” to the Israeli Harpy*. While this range of weapons systems have autonomous aspects they differ in the complexity of their algorithms and software, or their level of intelligence Landmines merely sense pressure, movement, sound, magnetism or vibration while the Harpy “detects, attacks and destroys enemy radar emitters, hitting them with high hit accuracy” [2]. The increasing complexity of weapons systems and their software led to the introduction of AI into the LAWS discussion. International discussions about lethal autonomous weapons systems (LAWS) have become more prevalent as nations attempt to come to terms with how these weapons work and how tc regulate them. The terms in which the community discusses them can at times be ambiguous as there is an absence of a common vocabulary and common understanding of the terms autonomy and artificial intelligence (AI). A weapon system is comprised of four actions, searching for < target, detecting that target, making the decision to engage the target, and engaging the target. [1 Developed by Colonel John Boyd, the OODA loop is a model for competitive decision making comprised of those four actions that define a weapon system: observe, orient, decide, and act. [1 In human operated weapons systems, a human operator completes all of these actions. In a weapor system with autonomy, the human operator relinquishes control over some part of the loop. autonomous as “freedom from external control or influence; independence.” Based on the above
Computer scientist John McCarthy first used the term artificial intelligence in 1955. [3 The term’s recent overwhelming use makes AI a point of discussion in many fields. Military anc civilian applications of AI are broad, including commuting, banking, farming, and search and rescue efforts. As a buzzword for political, academic, and private spheres, AI is frequently used as an umbrella term for fields including machine learning, data mining, speech recognition, anc image processing. By grouping such terms together, the term AI is frequently misused by technical experts, policymakers, and the general public. Figure 2 below shows Google’s search interests over the last five years for the terms autonomy and artificial intelligence. An overall upward trenc in the interest in AI is evident in the past three years. The international conversation on LAWS leaves several key issues unaddressed in the process of creating public policy regarding the use of autonomy and AI in weapons systems. A: advances in technology lead to increased autonomy and AI, the scope of these technologie: frequently overlaps, however it is a mistake to use the terms interchangeably because intelligence is not analogous with autonomy. A system does not need AI to be autonomous and the presence of AI does not imply the system is autonomous. To differentiate the terms, an understanding of the technical characteristics and bounds related to each is necessary. Another issue is < misunderstanding on how these technologies affect weapons systems. The lack of a universa definition of AI and autonomous weapon system leads to additional misunderstanding of the characteristics of AI and autonomy. Without comprehensive knowledge of the technologies policymakers cannot understand what impacts these technologies will have on LAWS. Thx absence of a common consensus within the international community on LAWS hinders the development of public policies, laws, and practices related to the use of autonomy and AI ir weapons systems.
Computer scientist John McCarthy first used the term artificial intelligence in 1955. [3 The term’s recent overwhelming use makes AI a point of discussion in many fields. Military anc civilian applications of AI are broad, including commuting, banking, farming, and search and rescue efforts. As a buzzword for political, academic, and private spheres, AI is frequently used as an umbrella term for fields including machine learning, data mining, speech recognition, anc image processing. By grouping such terms together, the term AI is frequently misused by technical experts, policymakers, and the general public. Figure 2 below shows Google’s search interests over the last five years for the terms autonomy and artificial intelligence. An overall upward trenc in the interest in AI is evident in the past three years. The international conversation on LAWS leaves several key issues unaddressed in the process of creating public policy regarding the use of autonomy and AI in weapons systems. A: advances in technology lead to increased autonomy and AI, the scope of these technologie: frequently overlaps, however it is a mistake to use the terms interchangeably because intelligence is not analogous with autonomy. A system does not need AI to be autonomous and the presence of AI does not imply the system is autonomous. To differentiate the terms, an understanding of the technical characteristics and bounds related to each is necessary. Another issue is < misunderstanding on how these technologies affect weapons systems. The lack of a universa definition of AI and autonomous weapon system leads to additional misunderstanding of the characteristics of AI and autonomy. Without comprehensive knowledge of the technologies policymakers cannot understand what impacts these technologies will have on LAWS. Thx absence of a common consensus within the international community on LAWS hinders the development of public policies, laws, and practices related to the use of autonomy and AI ir weapons systems.
— = aA ee ee oe ee A group of N GOs Teueichesd The Campaign to Stop Killer Robots, an international coalitior o ban fully-autonomous weapons in 2013. Participating organizations include Human Rights Watch, Amnesty International, Harvard Law School International Human Rights Clinic, and the International Committee for Robot Arms Control. The coalition’s main concern is that weapons will operate on their own without “meaningful human control”. The Campaign aims to inform solicymakers and the general public about the issues concerning LAWS, and makes statements and provides resources, such as the one shown below, to educate their audience.
— = aA ee ee oe ee A group of N GOs Teueichesd The Campaign to Stop Killer Robots, an international coalitior o ban fully-autonomous weapons in 2013. Participating organizations include Human Rights Watch, Amnesty International, Harvard Law School International Human Rights Clinic, and the International Committee for Robot Arms Control. The coalition’s main concern is that weapons will operate on their own without “meaningful human control”. The Campaign aims to inform solicymakers and the general public about the issues concerning LAWS, and makes statements and provides resources, such as the one shown below, to educate their audience.
Figure 5. Howard Gardener's 9 Types of Intelligence [8]
Figure 5. Howard Gardener's 9 Types of Intelligence [8]
Regarding AI, discussions of intelligence raise two additional questions, what 1s intelligence in machines...for example, is a machine intelligent if the system lacks linguistic intelligence and therefore cannot verbally express what it means? How do we measure a machine’ s intelligence? Alan Turing, in 1950, tackled this question and developed the Turing Test, one of the most well-known measures for machine intelligence. Turing designed the test to determine if < computer is capable of thinking like a human being. Below is an overview of the Turing Test. Automation v. Autonomy & Autonomous
Regarding AI, discussions of intelligence raise two additional questions, what 1s intelligence in machines...for example, is a machine intelligent if the system lacks linguistic intelligence and therefore cannot verbally express what it means? How do we measure a machine’ s intelligence? Alan Turing, in 1950, tackled this question and developed the Turing Test, one of the most well-known measures for machine intelligence. Turing designed the test to determine if < computer is capable of thinking like a human being. Below is an overview of the Turing Test. Automation v. Autonomy & Autonomous
Fully-autonomous, intelligent weapons systems do not exist yet. The “killer robot” “Terminator” is not yet a reality. Predicting exactly what the effects of these technologies will b on weapons systems is difficult. Some experts argue that the potential for LAWS to improve th accuracy and reduce the civilian casualties in war means the international community should nc pan them. They might have some positive functions in the future, with proper regulation an international discussion. If that is the case, some scholars argue that banning the technology an its development would be irresponsible. Implementing the technology that will cause the lea: amount of damage in international conflict would be our responsibility. Does restricting them no\ possibly limit to what extent we can use AI and autonomy to improve conflicts for both warfighte and civilian? Supporters of a ban want to ban fully-autonomous systems, but what about currer semi-autonomous systems such as missile defense systems, many of which engage a target withot he direct input of the human operator. Banning a technology so widely used by militaries acros the world may prove difficult, but not an impossible task, as exemplified by the ban on anti personnel landmines as well as other weapons.
Fully-autonomous, intelligent weapons systems do not exist yet. The “killer robot” “Terminator” is not yet a reality. Predicting exactly what the effects of these technologies will b on weapons systems is difficult. Some experts argue that the potential for LAWS to improve th accuracy and reduce the civilian casualties in war means the international community should nc pan them. They might have some positive functions in the future, with proper regulation an international discussion. If that is the case, some scholars argue that banning the technology an its development would be irresponsible. Implementing the technology that will cause the lea: amount of damage in international conflict would be our responsibility. Does restricting them no\ possibly limit to what extent we can use AI and autonomy to improve conflicts for both warfighte and civilian? Supporters of a ban want to ban fully-autonomous systems, but what about currer semi-autonomous systems such as missile defense systems, many of which engage a target withot he direct input of the human operator. Banning a technology so widely used by militaries acros the world may prove difficult, but not an impossible task, as exemplified by the ban on anti personnel landmines as well as other weapons.
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved