What are the ethical and legal challenges of Autonomous Weapon Systems in modern warfare?

Autonomous Weapon Systems (AWS) are weapons that can select and engage targets after activation without further human intervention. The central concern is not simply whether a machine can make a targeting decision, but whether people can ensure that decision complies with the laws and principles governing armed conflict.

Ethical challenges

A major concern is the loss of meaningful human control over decisions to use lethal force. A system may identify a person or object as a target, but it cannot understand human suffering or weigh moral responsibility as a person does. That raises difficult questions about whether delegating such decisions is acceptable, even if a system performs accurately in some conditions.

Reliability is another issue. Sensors and algorithms can misclassify people or objects, especially in cluttered environments or when conditions differ from those used in testing. Bias in training data, spoofing, hacking, and unexpected interactions with other systems can also produce harmful outcomes. Autonomous targeting decisions could further make conflict easier to initiate by reducing the perceived risk to a state’s own forces.

Legal challenges

Existing international humanitarian law applies to the use of weapons, including rules requiring distinction between civilians and combatants, proportionality, and precautions to protect civilians. The challenge is establishing whether a particular AWS can meet those rules reliably in the circumstances where it will be used. A system that performs acceptably in a controlled test may not be predictable enough for a changing battlefield.

Accountability can be difficult to trace when a system causes an unlawful death. Responsibility may involve commanders who deploy it, operators who set its parameters, designers, manufacturers, or the state that uses it. International law already provides routes for state responsibility and individual criminal liability, but proving who knew what, and whether a harmful result was foreseeable, can be complicated when decisions are partly automated.

There is no comprehensive global treaty that specifically bans or regulates every autonomous weapon. Discussions at the United Nations Convention on Certain Conventional Weapons have focused on possible limits and requirements, including human control, but states have not reached a universally binding agreement. For now, the debate is about how existing law should be applied and whether new rules are needed to ensure that humans remain meaningfully responsible for the use of force.