What does Preamble of Secure and Accountable Military AI Act of 2026 do?

This bill is formally titled 'To provide for secure and accountable use of artificial intelligence by the Department of Defense, and for other purposes.'

What does SHORT TITLE. of Secure and Accountable Military AI Act of 2026 do?

This bill is formally titled the Secure and Accountable Military AI Act of 2026.

What does DEFINITIONS. of Secure and Accountable Military AI Act of 2026 do?

This section defines key terms used throughout the Act. “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 select and engage targets without further intervention. “Covered contract” refers to agreements involving the development or provision of artificial intelligence models or systems for the Department of Defense. “Covered incident” includes theft of model weights, attempts by foreign adversaries to access sensitive information, supply chain compromises, data corruption, and concerning model behavior. “Frontier artificial intelligence model” is designated by the Secretary as presenting significant national security relevance. “High-consequence artificial intelligence application” refers to uses of AI by the Department presenting heightened risk. “Local defense” is defined as defense within a specific geographic area or of a high-value asset. “Operational deployment” means use of an AI system in support of a military mission. “Operator-supervised autonomous weapon system” is an autonomous weapon system allowing operator intervention.

What does HIGH-CONSEQUENCE MILITARY ARTIFICIAL INTELLIGENCE OVERSIGHT. of Secure and Accountable Military AI Act of 2026 do?

This section establishes oversight requirements for high-consequence military artificial intelligence (AI) applications within the Department. The Secretary must designate categories of AI applications as high-consequence within 180 days of the Act's enactment. Specific categories designated as high-consequence include AI used for nuclear weapon targeting, lethal targeting decisions, cyber operations with external effects, autonomous weapon systems, and domestic person-based analysis. Operational deployment of high-consequence AI applications requires written approval from a senior Department official. Before approval, the Department must complete testing, legal and policy reviews, and documentation of various aspects of the application, including limitations and fallback procedures. The Secretary must notify Congress 15 days before initial operational deployment, unless extraordinary national security circumstances exist, in which case notification can be delayed to 48 hours. The Secretary must provide an annual briefing to the Committees on Armed Services regarding the implementation of this section.

What does HUMAN ACCOUNTABILITY FOR HIGH-CONSEQUENCE MILITARY ARTIFICIAL of Secure and Accountable Military AI Act of 2026 do?

This section establishes Department policy regarding the use of artificial intelligence (AI) in high-consequence applications. AI may support decision-making but cannot replace accountable human judgment in areas like the use of force or detention. For each approved high-consequence AI application, the Secretary must ensure a clearly identified accountable human decision-maker or chain of decision-makers. Accountable decision-makers must have documented commander intent, training on the AI's capabilities and limitations, and access to relevant information. They must also have the authority to pause, override, deactivate, or terminate the AI's use and procedures for escalating issues.

What does REPORTING OF SECURITY INCIDENTS AND CONCERNING MODEL BEHAVIORS of Secure and Accountable Military AI Act of 2026 do?

This section establishes requirements for reporting security incidents related to artificial intelligence (AI) models and systems developed by contractors. The Secretary is required to include a reporting clause in covered contracts, either upon initial entry or within 90 days of the Act's enactment. The Secretary must issue guidance to contractors within 90 days of enactment, including examples of reportable incidents, reporting procedures, and timelines. Covered contractors must report incidents within 72 hours for certain events and 7 days for others, based on severity and potential impact. Reports must include detailed information about the incident, including its description, date, affected AI components, means of compromise, and potential impact. The Secretary is responsible for transmitting reports to various Department and Federal entities. The Department must notify Congress within 7 days of receiving an initial report or discovering an incident, providing regular updates as needed. Procedures will be established to protect sensitive information contained within the reports, including classified and proprietary data.

What does LIMITATIONS ON CERTAIN USES OF ARTIFICIAL INTELLIGENCE BY THE of Secure and Accountable Military AI Act of 2026 do?

This section establishes limitations on the Department's use of artificial intelligence (AI). The Department is prohibited from using AI for selecting nuclear weapon targets or executing launches. Generally, the Department cannot use AI to acquire or analyze information concerning U.S. persons within the U.S. for specific purposes like identifying, tracking, or assigning risk scores, unless expressly authorized by federal law. The Department is restricted from developing or employing autonomous weapon systems, with certain exceptions. Exceptions to the AI limitations exist for cybersecurity, force protection, and with informed consent. The Department cannot use AI based solely on First Amendment-protected activity or to circumvent legal protections by using publicly available or hacked data. Specific exceptions are made for semi-autonomous and operator-supervised autonomous weapon systems under certain conditions, such as defense against time-critical attacks. Any excepted autonomous weapon system must meet specific conditions, including legal review, rigorous testing, and human oversight. The Secretary must certify compliance with these conditions and notify Congress before fielding any system falling under the exceptions.

What does IMPLEMENTATION. of Secure and Accountable Military AI Act of 2026 do?

This section outlines the implementation procedures for the Act. The Secretary is required to revise the Defense Federal Acquisition Regulation Supplement and issue necessary regulations, guidance, and instructions within 180 days of the Act's enactment. The Secretary must provide a briefing to the Senate and House Committees on Armed Services within 240 days of enactment. The briefing must cover planned regulatory changes, directive and instruction updates, and potential resource or organizational issues affecting implementation.

What does AUTHORIZATION FOR SPECIALIZED AUTONOMOUS WEAPON SYSTEMS. of Secure and Accountable Military AI Act of 2026 do?

This section outlines the process for authorizing specialized autonomous weapon systems. The Secretary of Defense can submit a request for authorization to the House and Senate Armed Services Committees. Requests must include a classified annex detailing military necessity, operational constraints, law of armed conflict compliance, a risk mitigation plan, and a human judgment framework. Authorization requires a joint resolution of approval specifically naming the system and its authorized mission set. Congress has 20 calendar days to consider the joint resolution; failure to act results in automatic referral to the calendar. Senate procedures include a privileged motion to proceed, limits on debate (10 hours total), and prohibit amendments. Authorizations expire three years after enactment, with the possibility of renewal requests. The Secretary must certify that the system has operated within approved parameters for renewal.

Back to bills

Secure and Accountable Military AI Act of 2026

Last updated: 6/2/2026 · Introduced: 6/2/2026

Author: Kirsten Gillibrand (D-NY)

TL;DR (AI)

The bill establishes oversight requirements for high-consequence military artificial intelligence (AI) applications, mandating written approval from a senior Department official and notifying Congress before operational deployment.
It mandates that AI may support decision-making but cannot replace accountable human judgment, particularly in areas involving the use of force or detention.
The bill requires contractors to report security incidents and concerning model behaviors within specific timeframes, with the Department responsible for transmitting these reports to relevant entities and notifying Congress.
The Department of Defense is prohibited from using AI to select nuclear weapon targets or analyze information concerning U.S. persons within the U.S. without specific legal authorization.
The bill creates a process for the Secretary of Defense to request congressional authorization for specialized autonomous weapon systems, requiring a detailed submission and a joint resolution of approval.

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 7 more.

9 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