Act of the World Parliament on the Regulation of Autonomous Weapon Systems

Preamble

The World Parliament,

Affirming the imperative to uphold and strengthen International Humanitarian Law (IHL) and International Human Rights Law (IHRL), as well as the principles of humanity and the dictates of public conscience;

Recognizing the rapid advancements in artificial intelligence, robotics, and autonomy, and their potential application in weapon systems, which necessitate a robust and forward-looking international legal framework;

Concerned by the profound ethical, legal, humanitarian, and security implications posed by Autonomous Weapon Systems (AWS) operating without meaningful human control, including the risk of exacerbating conflicts, lowering the threshold for armed conflict, and undermining human agency and accountability;

Emphasizing the fundamental importance of maintaining human control over the use of force and ensuring human accountability for decisions concerning life and death;

Acknowledging the need for a preventive approach to address the risks associated with AWS, while also recognizing the potential for responsible innovation within a clearly defined regulatory framework;

Determined to establish clear international norms and regulations to prevent the development and deployment of AWS that pose unacceptable risks to humanity and stability, and to ensure that any permissible AWS operate under strict human oversight;

Hereby adopts this Act:

*

Chapter I: General Provisions

Article 1: Title and Scope

1. This Act shall be known as the "Act on the Regulation of Autonomous Weapon Systems."
2. This Act establishes a comprehensive international legal framework for the regulation, prohibition, and oversight of Autonomous Weapon Systems (AWS) to ensure compliance with international law and ethical principles.

Article 2: Definitions

For the purposes of this Act:

1. Autonomous Weapon System (AWS) means a weapon system that, once activated, can select and engage targets without further human intervention.
2. Meaningful Human Control (MHC) means a level of human involvement and oversight in the operation of an AWS sufficient to ensure compliance with International Humanitarian Law (IHL) and International Human Rights Law (IHRL), maintain human agency and accountability, and allow for human judgment in critical decisions. This includes the ability to intervene, deactivate, or override the system in a timely and effective manner.
3. Critical Functions refer to the selection and engagement of targets by a weapon system.
4. Human-in-the-Loop refers to an AWS that requires human authorization for each individual engagement decision.
5. Human-on-the-Loop refers to an AWS that allows humans to monitor the system's target selection and engagement decisions and to intervene or override them before engagement.
6. Human-out-of-the-Loop refers to an AWS that, once activated, operates without human intervention in the critical functions of target selection and engagement.

*

Chapter II: Core Principles and Prohibitions

Article 3: Principle of Meaningful Human Control

All Autonomous Weapon Systems shall be designed, developed, tested, deployed, and used under meaningful human control over their critical functions.

Article 4: Prohibition of Certain Autonomous Weapon Systems

The following Autonomous Weapon Systems are hereby prohibited:

1. AWS that are designed or deployed to operate without meaningful human control over their critical functions.
2. AWS that are inherently incapable of complying with International Humanitarian Law (IHL), including the principles of distinction, proportionality, and precaution in attack.
3. AWS that are inherently indiscriminate or unpredictable in their effects, thereby posing an unacceptable risk to civilians and civilian objects.
4. AWS that are designed or intended to target human beings directly without human intervention in the final decision to engage.
5. AWS that are designed or intended to cause widespread, long-term, and severe damage to the natural environment.

Article 5: Accountability

Human accountability for the development, deployment, and use of all AWS, including for any unlawful acts committed by or through such systems, shall be ensured at all times. The absence of a clear chain of human command and control shall not absolve States or individuals of responsibility.

*

Chapter III: Regulatory Framework for Permissible AWS

Article 6: Design and Development Standards

Member States shall ensure that any permissible AWS are designed and developed in accordance with the following standards:

1. Transparency and Explainability: The design and underlying algorithms, data sets, and decision-making processes of AWS shall be sufficiently transparent and explainable to allow for human understanding, predictability, and accountability.
2. Predictability and Reliability: AWS shall be designed to operate predictably and reliably, with clearly defined operational parameters and limitations, and shall incorporate robust mechanisms to prevent unintended or unauthorized operations.
3. Built-in Safeguards: AWS shall incorporate built-in safeguards, including clear human-machine interfaces, easy deactivation mechanisms, human override capabilities, and fail-safe modes that revert to human control or cease operation in the event of system failure or unforeseen circumstances.
4. Testing and Validation: Rigorous and comprehensive testing and validation protocols shall be mandatory throughout the entire lifecycle of an AWS, including simulated and real-world scenarios, to assess performance, reliability, and compliance with IHL and IHRL.
5. Ethical Considerations: Ethical implications, including the potential for bias in data sets or algorithms, shall be thoroughly assessed and addressed during the design and development phases.

Article 7: Deployment and Use Restrictions

Member States shall ensure that any permissible AWS are deployed and used under strict restrictions, including:

1. Rules of Engagement: Clear and specific rules of engagement that ensure meaningful human control over critical functions, requiring human review and authorization for target engagement decisions.
2. Operational Context: Prohibition of deployment in environments where the system's predictability, reliability, or ability to distinguish between combatants and civilians cannot be guaranteed.
3. Human Supervision: Continuous human supervision and monitoring of AWS operations, with the capability for immediate intervention or termination of operations.
4. Proportionality and Precaution: Operators shall remain responsible for conducting proportionality assessments and taking all feasible precautions to avoid, or in any event to minimize, incidental loss of civilian life, injury to civilians, and damage to civilian objects, in accordance with IHL.

Article 8: Responsibility and Liability

1. Member States shall establish clear national frameworks for attributing responsibility and liability for any unlawful acts or damages resulting from the development, deployment, or use of AWS.
2. Such frameworks shall ensure effective mechanisms for victim redress and compensation.

*

Chapter IV: International Cooperation and Oversight

Article 9: World Parliament Committee on Autonomous Weapon Systems (WPC-AWS)

1. A standing World Parliament Committee on Autonomous Weapon Systems (WPC-AWS) is hereby established to:
a. Monitor compliance with this Act by Member States.
b. Develop and update technical guidelines and best practices for the design, development, testing, and deployment of AWS.
c. Facilitate international cooperation and information exchange among Member States on AWS-related matters.
d. Conduct independent assessments and provide expert advice to the World Parliament on emerging technological developments and their implications for this Act.
e. Promote universal adherence to this Act.

Article 10: Transparency and Reporting

Member States shall submit annual reports to the WPC-AWS detailing their policies, research, development, and deployment of AWS, including measures taken to ensure compliance with this Act. These reports shall be made publicly available, subject to legitimate national security concerns.

Article 11: Capacity Building

Member States shall cooperate to provide assistance, including technical and financial support, to other Member States to facilitate the effective implementation of this Act.

*

Chapter V: Implementation and Review

Article 12: National Implementation

Member States shall take all necessary legislative, administrative, and other measures to ensure the full and effective implementation of this Act within their national jurisdictions.

Article 13: Review Conference

1. A Review Conference of the States Parties to this Act shall be convened by the World Parliament every five years, or earlier if deemed necessary, to:
a. Assess the operation and effectiveness of this Act.
b. Consider any amendments to this Act in light of new technological developments, evolving ethical considerations, and experiences gained from its implementation.
c. Review the progress towards achieving the objectives of this Act.

Article 14: Entry into Force

This Act shall enter into force six months after the date of deposit of the [Number, e.g., twentieth] instrument of ratification, acceptance, approval, or accession with the Secretariat of the World Parliament.