斯德哥尔摩国际和平研究所-Mapping-the-Development-of-Autonomy-in-Weapon-Systems_147页_6mb

Summary of Mapping the Development of Autonomy in Weapon Systems

Core Content

This report, conducted by the Stockholm International Peace Research Institute (SIPRI) in 2017, presents a comprehensive mapping study on the development of autonomy in weapon systems. It aims to support discussions under the Convention on Certain Conventional Weapons (CCW) and provide insights into the technological, operational, strategic, and ethical implications of increasing autonomy in military systems.

Main Findings

1. Technological Foundations of Autonomy

• Definition of Autonomy: Autonomy refers to the ability of a machine to perform an intended task without human intervention, using sensor data and computer programming to interact with the environment.
• Key Technologies: Autonomy primarily relies on software, with machine learning and artificial intelligence (AI) playing a significant role in its development. The feasibility of autonomy depends on:
  • The ability to formulate tasks as mathematical problems.
  • The capacity to model or map the operating environment in advance.
• Autonomy in Weapon Systems: Autonomy is not a single feature but a complex system of functions. It includes reactive and deliberative capabilities, with the latter involving more advanced decision-making processes.
• Challenges: The use of machine learning in weapon systems remains experimental due to fundamental issues with predictability. Current systems are limited in perceptual and decision-making intelligence, especially in cluttered or adverse environments.

2. State of Autonomy in Weapon Systems

• Existing Capabilities: Autonomy is already used in various weapon systems, including mobility, targeting, intelligence, interoperability, and health management.
• Automated Target Recognition (ATR): ATR systems, which enable autonomous target acquisition, have been in use since the 1970s but are still limited in performance.
• Defensive Systems: Most existing weapon systems with autonomous targeting capabilities are defensive and operate under human supervision. They are designed to engage targets autonomously only in time-critical situations.
• Offensive Systems: Loitering weapons are the only known offensive weapon systems capable of autonomous target acquisition and engagement. Their deployment parameters and target categories are defined in advance by humans.

3. Drivers and Obstacles to Autonomy Development

• Strategic Drivers: The U.S. has emphasized autonomy as a cornerstone of its military modernization and strategic capabilities. This has prompted similar interest from Russia and China.

• Operational Drivers: Military planners believe autonomy enhances speed, accuracy, persistence, reach, and coordination on the battlefield.

• Economic Drivers: Autonomy is seen as a way to reduce operating costs through more efficient use of manpower.

• Technological Obstacles: Autonomous systems must be more adaptive to operate safely in complex, dynamic, and adversarial environments. New validation and verification methods are needed for systems that can learn or adapt.

• Institutional Resistance: Military personnel often lack trust in autonomous systems, and some professionals view autonomy as incompatible with traditional operational paradigms.

• Legal Obstacles: International law requires human control or oversight over weapon systems, limiting the use of fully autonomous targeting capabilities.

• Normative Obstacles: Civil society increasingly opposes the use of autonomy in targeting decisions, raising ethical and political concerns.

• Economic Obstacles: National armed forces face affordability challenges, and defense acquisition systems are not well-suited to support the development of autonomous technologies.

4. Innovation Ecosystems

• Key Disciplines: Advances in autonomy are driven by research in artificial intelligence (AI), robotics, and control theory.
• Leading Nations: The U.S. leads in military R&D on autonomy, while China and other major arms-producing countries are also investing heavily in AI and robotics.
• Civilian Industry: The civilian sector, particularly in information technology and automotive industries (e.g., Alphabet, Amazon, Baidu, Toyota), is a major driver of innovation in autonomous technologies.
• Defence Industry: While traditional arms producers are involved, they allocate fewer resources to autonomy R&D compared to commercial entities. However, their role remains crucial as they often modify civilian technologies for military use.

Key Recommendations

The report concludes with eight recommendations for future discussions on Lethal Autonomous Weapon Systems (LAWS) within the CCW framework:

1. Focus on the development of autonomy in weapon systems rather than on LAWS as a general category.
2. Shift the focus from full autonomy to how autonomy transforms human control.
3. Expand the scope of investigation beyond targeting to include collaborative operations and intelligence processing.
4. Clarify the implications of machine learning on the control of autonomous systems.
5. Use case studies to connect discussions on legality, ethics, and human control with real-world applications.
6. Facilitate knowledge exchange with the civilian sector on definitions, human control, and validation of autonomous systems.
7. Ensure that monitoring and regulation of LAWS do not hinder civilian innovation.
8. Explore ways to prevent the weaponization of civilian technologies by non-state actors.

Key Information

• The study analyzed 381 weapon systems and unmanned military robotic systems from 12 countries.
• Autonomy is a multifaceted concept with varying interpretations across disciplines.
• The U.S. has been the most active in military R&D on autonomy, followed by China and other major arms-producing nations.
• Civilian innovation is critical to the development of autonomous technologies, though military adaptation often requires modifications.
• Legal and ethical concerns are significant, especially regarding the role of humans in decision-making processes.

Conclusion

This report provides a detailed mapping of the development of autonomy in weapon systems, highlighting the technological, strategic, and ethical dimensions of its advancement. It serves as a foundational resource for ongoing discussions on the regulation of autonomous weapon systems under the CCW framework.