Artificial Intelligence in Future Warfare: Ethical Frameworks and the Regulation of Lethal Autonomous Weapons

The integration of artificial intelligence into weapons systems has compressed the decision cycle of lethal engagement
from hours to milliseconds, outpacing the international legal and ethical frameworks designed to constrain state
violence. This paper surveys the landscape of deployed and tested lethal autonomous weapons systems (LAWS),
analyzes the adequacy of existing international law relative to current AI capabilities, and proposes a regulatory
structure calibrated to the actual risk profile of autonomous lethality. We examine nine real-world systems -- from the
Kargu-2's documented autonomous engagement in Libya (2020) to Israel's "Lavender" AI targeting in Gaza
(2023-2024) and the ongoing 2026 Iran-US-Israel conflict "Operation Epic Fury," the largest AI-assisted warfare
campaign in recorded history -- and classify each using a three-tier autonomy model: human-in-the-loop (HITL),
human-on-the-loop (HOTL), and human-out-of-the-loop (HOOTL). Our gap analysis of the Geneva Conventions, the
Convention on Certain Conventional Weapons (CCW), and International Humanitarian Law (IHL) identifies four critical
regulatory failures: the absence of a binding definition of "meaningful human control," an accountability vacuum when
LAWS cause civilian casualties, a speed asymmetry between AI warfare timescales and legal review processes, and
the dual-use nature of civilian AI technologies. To address these gaps, we propose a five-tier governance framework
scaling regulatory stringency with the product of autonomy level and lethality threshold. The framework carries direct
implications for stalled UN CCW Group of Governmental Experts negotiations, offering a technically grounded basis for
legally binding distinctions that current diplomatic language lacks.