Source: Lieber Institute West Point

An article from the Lieber Institute for Law & Land Warfare at West Point provides a rigorous examination of the legal review process that the U.S. Department of Defense (DoD) applies to new military capabilities incorporating artificial intelligence and autonomy. Authored by a legal expert in the field, the piece centers on DoD Directive 3000.09, which mandates that all autonomous and semi-autonomous weapon systems undergo a thorough legal review to ensure they function in accordance with the laws of war, also known as international humanitarian law. This review process is critical for maintaining legal and ethical standards as the military increasingly relies on advanced technologies.

The author, Paul M. Bailey, highlights the profound challenges that AI presents to these established review mechanisms. A key difficulty is the "black box" problem, where the decision-making processes of complex machine learning models can be opaque even to their designers, making it hard to predict and verify their compliance with principles like distinction, proportionality, and precaution. The article stresses the importance of continuous and rigorous testing, evaluation, and validation throughout the AI system's lifecycle, especially as the system learns and adapts over time. It concludes that a robust, multi-disciplinary legal and technical review process is essential to guarantee that "meaningful human control" is maintained and that the deployment of AI in warfare remains within the bounds of international law.