Editors’ note: This post is part of a series featuring topics discussed during the Third Annual Future of Warfare and the Law Symposium. Christina Colclough’s introductory post is available here.

The Future of Warfare and the Law Symposium (hereinafter, the “Symposium”), which took place in May 2025, brought together lawyers and technical experts to discuss some of the most pressing challenges concerning the future of warfare, including military uses of artificial intelligence (AI). This post addresses a topic that generated significant discussion in the workshop: how the need to rapidly modify autonomous weapons to counter enemy deception intersects with the legal obligation to review those modifications.

During the Symposium, participants explored a scenario in which an adversary employed deception tactics to evade autonomous targeting systems. In the Symposium example, the enemy placed tires on tanks, causing friendly force sensors to fail to recognize them as military objectives. The tactic provoked a software update, which raised a broader concern: how can friendly forces quickly adapt when adversaries continuously and rapidly change their deception methods? To remain effective, belligerents should update weapons using artificial intelligence (AI)/machine learning (ML) targeting quickly in the field, sometimes in hours, not weeks or months. At what point do these adjustments require a new legal review? How does the current weapon review framework account for such rapid change by opposing forces, and how might it need to evolve to remain effective?

With those questions in mind, we turn to ongoing operations in Ukraine, where opposing forces are engaged in a persistent, competitive cycle of rapid battlefield innovation. In an iterative process, today’s innovation by one side leads to rapid innovation on the other side. The pace of tactical innovation demonstrated on the forward line of troops in Ukraine represents a trend that generally strains current U.S. policies and doctrine regarding weapons reviews.

Further, as Ukraine and Russia increasingly innovate toward autonomous weapon systems, the scenario discussed at the Symposium is closer to reality. The United States should not wait until this cycle of rapid innovation meets the proliferation of autonomous systems on the future battlefield before considering modifications to existing policies.

Legal Reviews for Weapon Systems

We now turn to the requirements under international law for the legal review of weapons. The clearest articulation of such an obligation is found in Article 36 of Additional Protocol I (AP I), which obliges parties to determine whether the use of a new weapon, means, or method of warfare would be prohibited under any rule of international law. Although the United States is not a party to AP I, and Article 36 is not generally regarded as customary international law, the obligation to review weapons does not rest solely on that provision. Nonetheless, U.S. policy and practice of conducting legal review of weapons predates AP I.

Legal review requirements may also exist in other rules of the law of armed conflict (LOAC). States have committed to respect and ensure respect for LOAC, to provide their armed forces with instruction on its application, and to refrain from employing unlawful weapons. Collectively, these commitments imply a responsibility to assess new weapons, technologies, and methods of warfare.

Legal advisers play a central role in this process. Article 82 of AP I requires that legal advisers be available to military commanders to advise on the application of the Geneva Conventions and the Additional Protocols, as well as on the training of the armed forces in these rules. Importantly, while the United States has objected to many provisions of Additional Protocol I, it has not expressed dissent to Article 82.

Current U.S. policy sets clear requirements for reviewing weapons for legal sufficiency. DoD Directive 2311.01 (Law of War Program) requires all new or modified weapons to undergo a legal review. DoD Directive 5000.01 (Defense Acquisition System) provides that the acquisition and procurement of DoD weapons and information systems must be consistent with all applicable domestic law, and the resulting systems must comply with applicable treaties and international agreements, customary international law, and the law of armed conflict. It further mandates that an authorized attorney conduct the review.

These reviews focus on three questions expressed in the Department of Defense Law of War Manual: 1) whether a weapon is designed to cause unnecessary suffering; 2) whether it is inherently indiscriminate; and 3) whether it falls into a prohibited category.

If a weapon clears these thresholds, attorneys also assess any restrictions on its use, evaluate the proposed concept of employment, and may recommend additional measures such as training, doctrine, or rules of engagement to ensure compliance. For autonomous and semi-autonomous systems, DoD Directive 3000.09 reiterates that legal reviews for intended acquisition or modification of autonomous weapons are conducted in accordance with these regulations.

Additionally, current Army policy requires legal review of all weapons and weapon systems under Army Regulation 27-53, Legal Review of Weapons and Weapons Systems. The U.S. Army Judge Advocate General’s National Security Law Division (NSLD) conducts these reviews on behalf of the Judge Advocate General (TJAG), though TJAG may appoint other offices as designees for specific systems. Legal reviews are necessary before any weapon is fielded. When Soldiers modify a fielded system in a way that substantially changes its intended use, or anticipated effects, they invalidate the original legal opinion, requiring a new review by NSLD.

The Way Forward

Current events in Ukraine underscore two primary issues with current policies. First, the centralized process for generating legal reviews when substantive changes trigger the need for a new review creates a single point of failure that cannot possibly keep pace with the speed of tactical innovation. Second, modern technology, including autonomous programming, has made innovation at the tactical edge easier than ever. As autonomous weapons capabilities push closer to the tactical edge, the Army’s current centralized legal review process risks creating delays that could hinder the operational effectiveness of troops rapidly innovating to fight and win. To preserve both agility and compliance with international law, the Army should consider updating its approach to reviewing weapons. To this end, the Army should consider at least three potential solutions.

First, the Army can decentralize initial review. This would empower unit legal advisors to conduct preliminary reviews, with final approval routed upward through the legal technical chain as the weapon systems mature. Moreover, decentralization affords an opportunity to connect legal reviews to acquisition authorities during peacetime (e.g., a brigade judge advocate can draft a legal review during experimentation, with the final review reserved for NSLD post-acquisition), with enhanced delegations to lower approval levels during named combat operations.

Second, the Army can delegate certain categories of legal reviews to theater-level commanders and legal staff, reducing reliance on headquarters-level approvals. One potential variant here would include an overarching legal review over a base system or capability, with the theater empowered to review modifications to ensure continued compliance with LOAC. An additional consideration would involve more strictly defining what constitutes a substantial modification to a weapons platform. For example, the Army should consider at what point retraining a previously reviewed (and legally sound) autonomous system on new enemy tactics, techniques, and procedures requires a new legal review.

Finally, the Army can modernize the traditional weapons review process by establishing a streamlined review board at the NSLD level that has authority to set priorities and fast-track critical cases. This process could be automated using existing capabilities like SharePoint and powered by AI-enabled screening criteria to screen and prioritize legal reviews.

Concluding Thoughts

The United States has long been a leader in legal reviews of weapons. Its policies and practices have gone well beyond what is legally required. This system has served the force and its service members exceedingly well to date. However, the legacy system of weapon reviews requires adjustment to keep pace with the demands of an evolving security environment and dynamic weapon systems including cyber tools, AI and ML systems. The proposals of this post will retain the advantages of the legacy system, set a model for the evolution and interpretation of international legal standards, all while permitting the force to keep pace with pressing national security needs.

***

Colonel Andrew McKee currently serves as the Judge Advocate for United States Army Europe-Africa.

LTC Christina Colclough is an assistant professor in the Department of Law and Philosophy at the United States Military Academy.

The views expressed are those of the authors, and do not necessarily reflect the official position of the United States Military Academy, Department of the Army, or Department of Defense. 

Articles of War is a forum for professionals to share opinions and cultivate ideas. Articles of War does not screen articles to fit a particular editorial agenda, nor endorse or advocate material that is published. Authorship does not indicate affiliation with Articles of War, the Lieber Institute, or the United States Military Academy West Point.

 

 

 

 

 

 

Photo credit: U.S. Air Force, Master Sgt. Stanley Thompson