What does Preamble of HALO Act of 2026 do?

This bill, formally titled 'To provide for design and safety requirements for autonomous and semi-autonomous weapon systems, and for other purposes,' aims to establish design and safety requirements for autonomous and semi-autonomous weapon systems.

What does SHORT TITLE. of HALO Act of 2026 do?

This bill is formally titled the Human Authority in Lethal Operations Act of 2026.

What does DEFINITIONS. of HALO Act of 2026 do?

This section defines key terms used throughout the Act. “Appropriate committees of Congress” are specified as certain committees within the Senate and House of Representatives. “Artificial intelligence” is defined by reference to section 5002 of the National Artificial Intelligence Initiative Act of 2020. An “autonomous weapon system” is defined as one that can identify, select, or engage targets without further human intervention. A “covered artificial intelligence capability” refers to AI used by or shared with the Department of Defense. “Ethical Principles for Artificial Intelligence” are defined as those adopted by the Department on February 24, 2020, as in effect on January 1, 2025. “Responsible Artificial Intelligence Strategy and Implementation Pathway” is defined as a document prepared by the Department of Defense Responsible Artificial Intelligence Working Council. “Semi-autonomous weapon system” is defined as a system intended to only engage targets previously selected by a human operator, even if it autonomously conducts engagement-related functions. “Specific target group” is defined as a discrete group of potential targets, excluding general classes or types of targets. “Unintended engagement” refers to the use of force resulting in damage to persons or objects that human operators did not intend to target.

What does DESIGN AND SAFETY REQUIREMENTS FOR AUTONOMOUS AND SEMI- of HALO Act of 2026 do?

This section establishes design and safety requirements for autonomous and semi-autonomous weapon systems that use artificial intelligence to engage targets. The Secretary of each military department must ensure these systems meet the outlined requirements. The law mandates a clear chain of command and designates accountable commanders for engagements involving these systems, regardless of AI's contribution. Systems must incorporate design features that require designated commanders to exercise ultimate discretion and control, including constraints on authorized actions and scope. Records of target selection data, decision logic, and human operator actions must be maintained for post-engagement review. The Ethical Principles for Artificial Intelligence and the Responsible Artificial Intelligence Strategy and Implementation Pathway must be followed. Systems must be designed to complete engagements within designated parameters and undergo independent review before engaging previously unauthorized targets. The physical hardware and software of these systems must be auditable, explainable, and incorporate robust cybersecurity measures. Human operators must have clear procedures to activate, terminate, and disable weapon system functions, and receive timely feedback on system status. In the event of degraded communications, the system must not autonomously select and engage targets not previously authorized by a human operator. Continuous monitoring procedures are required to identify and address circumstances requiring additional testing or legal review.

What does ADVANCED REVIEW AND APPROVAL AUTHORITY. of HALO Act of 2026 do?

This section establishes an advanced review and approval authority for weapon systems described in section 3(a)(2). Systems generally require approval before formal development and fielding, with exceptions outlined in subsection (e). Before development, the Under Secretary of Defense for Policy, Research and Engineering, and the Vice Chairman of the Joint Chiefs of Staff must verify capabilities, design considerations, risk mitigation, verification plans, and legal analysis. Prior to fielding, the same officials must verify system capabilities, human-machine interfaces, doctrine, safety measures, training, and legal compliance. Modified systems with significant changes to algorithms, mission sets, or target sets require a new analysis and review. Certain semi-autonomous and human-supervised autonomous weapon systems are exempt from the advanced review process. However, systems used in densely populated areas or near essential civilian infrastructure, even if otherwise exempt, require advanced review to ensure the ability to distinguish between military and civilian targets.

What does ROLE OF THE CHIEF DIGITAL AND ARTIFICIAL INTELLIGENCE OFFICER. of HALO Act of 2026 do?

This section establishes the responsibilities of the Chief Digital and Artificial Intelligence Officer. The Officer shall monitor and evaluate artificial intelligence capabilities and cybersecurity for autonomous and semi-autonomous weapon systems, advising the Secretary on these matters. The Officer, in collaboration with the Under Secretary of Defense for Research and Engineering, will formulate testable requirements for implementing ethical AI principles and strategies. The Officer will establish policy and guidance on defining requirements and testability for AI-enabled systems to ensure adherence to ethical principles and strategies. Guidance will be issued on test and evaluation practices, including adversarial assessments (red-team assessments) to evaluate weapon system vulnerability. The Officer will coordinate with other officials to develop and use common tools and infrastructure for testing and verification of AI capabilities in weapon systems.

What does TESTING AND EVALUATION REQUIREMENTS. of HALO Act of 2026 do?

This section establishes testing and evaluation requirements for weapon systems described in section 3(a)(2). The Secretary must ensure rigorous hardware and software verification and validation, and realistic system testing, regardless of the acquisition pathway. Testing may include assessing human operator response time and ability to challenge system suggestions. Testing must verify system functionality in realistic operational environments against adaptive adversaries, including civilian presence. Systems incorporating artificial intelligence (AI) require rigorous testing to ensure AI robustness and the ability to rapidly reprogram autonomy algorithms. Post-fielding testing is required to ensure critical safety features are not degraded and to monitor system behavior. Quarterly cyber testing is mandated to verify system resilience and survivability in contested cyberspace. The Director of Operational Test and Evaluation (DOT&E) oversees the development of test standards and risk assessments for autonomous weapon systems. DOT&E is responsible for establishing data collection standards, maintaining a repository for incidents and failures, and reviewing test plans. The DOT&E coordinates with other departments to identify when additional testing is needed to address potential issues like unintended engagements with civilians.

What does PROHIBITED USES OF ARTIFICIAL INTELLIGENCE. of HALO Act of 2026 do?

This section prohibits the use of covered artificial intelligence (AI) capabilities in ways that violate the U.S. Constitution, federal law, the Law of Armed Conflict, or international legal obligations. The section also restricts AI use that poses an unacceptable risk to individual safety or civil liberties. Specific prohibitions include using AI to profile individuals based on constitutionally protected rights, detect emotional states, or infer sensitive personal information like religion, ethnicity, or political identity. Tracking an individual’s location without a judicial order or legal requirement is also prohibited. Aggregating data from various sources to achieve prohibited purposes is restricted. The section prevents removing humans from critical decision-making processes, particularly those related to nuclear weapons employment. Obtaining personal data from data brokers for a fee is generally prohibited. “Personal data” is defined as data linked to an individual or device, including anonymized data that can be re-linked. These prohibitions apply to all activities of the Department and extend to other federal agencies that receive AI systems, data, or analytical products from the Department. The restrictions also apply to Departmental participation in joint operations and interagency working groups.

What does WHISTLEBLOWER PROTECTIONS. of HALO Act of 2026 do?

This section requires the Secretary to update whistleblower protections to clarify procedures for artificial intelligence systems. These protections apply to personnel involved in the development, assessment, deployment, operation, or use of artificial intelligence within National Security Systems or for military/intelligence purposes. The updates must comply with existing laws, including sections 1034 and 4701 of title 10, section 2302 of title 5, and title VI of the Intelligence Authorization Act for Fiscal Year 2014. The Secretary must ensure procedures exist for receiving, investigating, and redressing anonymous complaints related to artificial intelligence. Confidentiality protections do not extend to significant misconduct, such as violations of law or government ethics. The Secretary must also ensure procedures exist for reporting incidents of artificial intelligence misuse, investigating them, and taking corrective actions.

What does REPORTING REQUIREMENTS. of HALO Act of 2026 do?

This section establishes reporting requirements for the Secretary regarding artificial intelligence (AI) use within the Department. The Secretary must submit a semiannual report on exemplary AI use cases, including best practices, failure modes, and risk mitigation strategies. After-action reports are required for significant operational use of AI, assessing performance, human oversight, risks, incidents, and recommending improvements. Training provided to human operators of AI systems must be described in the semiannual report, including employment and retirement procedures. An annual report is required to identify barriers to responsible AI development and deployment, as well as infrastructure gaps. The Secretary must also submit a semiannual report identifying fielded AI systems that do not comply with the Act, along with justifications and remediation plans.

Back to bills

HALO Act of 2026

Last updated: 6/8/2026 · Introduced: 6/8/2026

Author: Adam Schiff (D-CA)

TL;DR (AI)

The HALO Act establishes design and safety requirements for autonomous and semi-autonomous weapon systems, mandating a clear chain of command and requiring designated commanders to exercise ultimate control over engagements.
Weapon systems utilizing artificial intelligence generally require advanced review and approval before development and fielding, with exceptions for certain semi-autonomous systems and those used in densely populated areas.
The Chief Digital and Artificial Intelligence Officer is created to monitor AI capabilities, formulate testable requirements for ethical AI, and coordinate testing and verification efforts.
The Act mandates rigorous testing and evaluation of weapon systems, including assessments of human operator response, realistic operational environments, and cybersecurity resilience.
The HALO Act prohibits the use of artificial intelligence in ways that violate U.S. law, risk individual safety, or remove humans from critical decision-making processes, particularly concerning nuclear weapons.

119th Congressin committeeOfficial page

Verified Votes

No verified votes yet.

Community Votes

No community votes yet. Be the first!

Voting as guest · Sign in for verified votes

Other Sections

Includes provisions on preamble, short title., and 10 more.

12 sections

Click "Show details" to explore individual sections. Use the community vote above or visit All Sections to vote per section.

Read official text

Want the full experience?

Create a free account to cast verified votes on each section, get a personalized summary, and track your voting history.

Create Free Account