AI and Autonomous Weapons: The Ethics of Deterrence and Warfare
The Reckless Rush Toward Autonomous Weapons
In an AI arms race, states competing for supremacy and first-mover advantage will not pay adequate attention to safety and technical regulations.
This leap forward raises urgent questions about deterrence, accountability, and.
From Civil War Gatling Guns to Counterinsurgency Drones
Historically, automation served as the first foundational step toward full military autonomy.
During the American Civil War, Richard Gatling introduced an innovative loading and firing mechanism that enabled a single gun to fire 300 rounds per minute. That invention vastly outperformed rifled muskets, which managed roughly three rounds in the same timeframe.
Decades later, armed drones and early automated systems proved their utility on a practical scale during counterinsurgency conflicts in Iraq and Afghanistan. These deployments cemented the role of automated hardware in modern combat environments.
Proponents of autonomous weapon systems argue that these advanced technologies can ultimately make warfare more humane. Advocates maintain that machines can be programmed to strictly avoid civilian casualties, noting that algorithms do not tire, experience fear, or act out of rage and revenge. Supporters point to reduced human workloads and greater cost efficiency as driving factors behind the technological push.
Algorithmic Blind Spots and Technical Vulnerabilities
Critics of military artificial intelligence emphasize that ethical and moral considerations must outweigh cold agentic calculations.
Distinguishing between civilians and combatants demands complex contextual judgement, a capacity algorithms currently lack when tasked with calculating lethal force. Technical vulnerabilities compound these concerns. Agentic autonomous weapons systems can ingest corrupted data, exhibit severe operational misalignment, or be actively manipulated by hostile adversaries.
The pressure to secure a competitive edge creates a dangerous environment where states ignore essential safeguards.
Imagining a High-Stakes Crisis Without Human Judgment
Former Pentagon official Michael Horowitz highlighted these systemic risks by sketching an imaginary alternative to the Cuban missile crisis, asking how autonomous naval blockade ships would alter crisis stability.
Pre-programmed algorithms designed to automatically fire on blockade-runners might theoretically strengthen deterrence by placing the immediate onus of conflict on an adversary. Yet, severe communication gaps remain.
Moscow or Washington could easily dismiss an opponent’s deployment as a bluff, leading to unintended escalation.
Technical failures also introduce the terrifying prospect of accidental firing, leaving a critical vacuum regarding accountability when machines fail. While human operators may process information less efficiently than agentic machines, people retain the vital ability to grasp the broader strategic picture, understand an opponent’s perspective, and step back from the precipice before catastrophe strikes.