Autonomous weapons exist on a spectrum from semi-autonomous systems (where a human approves final action) to fully autonomous systems that operate from target identification to lethal engagement without any human input. 

At the centre of this debate is the challenge of human control; in particular, the concept of "meaningful human control" (MHC), which has become a defining feature in LAWS regulation, though as a term it presents significant hurdles due to its vagueness, lack of established practice, and absence of global consensus.

New fully autonomous (unmanned) weapons systems in development remove or reduce the human role in the control and decision-making process, with the goal of removing humans from the active battlefield en masse. The Pentagon's Replicator programme, for example, promises a drastic shift in warfare towards highly autonomous and cooperative AI units. 

Why it matters

Autonomous weapons challenge the most fundamental ethical and legal norms governing the use of lethal force. 

International humanitarian law (IHL) assumes human judgement as the locus of responsibility. Article 36 of Additional Protocol I to the Geneva Conventions implicitly presumes human oversight in weapons deployment, while the doctrine of mens rea in war crimes jurisprudence necessitates conscious intent. 

Autonomous weapons dissolve these foundations by introducing systems that process data without contextual moral comprehension.

At the heart of the debate lies the question of moral responsibility: if a machine independently makes a decision to kill, who is accountable - the programmer, the commander, the manufacturer, or the state?

Beyond individual conflicts, because highly capable LAWS lower the human costs associated with conflict initiation and escalation, they create a large risk to global geopolitical stability - and this effect worsens the more capable the systems become.

These systems challenge the principle of human dignity - the idea that a person should not be killed by a machine that cannot understand the value of life. Furthermore, they risk lowering the threshold for entering into war by reducing the physical risk to a nation's own soldiers. 

Impacts & consequences

• Accountability gap. If a machine commits a war crime, it is unclear who is held responsible: the programmer, the manufacturer, or the commanding officer. 

• Civilian harm. AI's built-in unpredictability becomes especially problematic in urban warfare, as systems may not adapt quickly to changing situations and may mistake civilians or civilian objects for enemies, leading to unintended IHL violations and escalation. 

• Conflict escalation. Autonomous weapons may make conflict more likely to escalate faster, since machines can operate at machine speed and reduce human deliberation. 

• Escalation unpredictability. AI-vs-AI "flash wars" could occur if two autonomous systems interact in unforeseen ways, leading to rapid escalation before humans can intervene. 

• Lower threshold for armed conflict. Reducing human casualties in a single conflict can be outweighed if the total number of conflicts increases. But the past century suggests that dominance in a new military technology does not always translate to greater stability. 

• Proliferation. Unlike nuclear weapons, AI software is easy to copy and distribute, making it accessible to non-state actors and rogue regimes. 

Causes & contributing factors

The use of autonomous weapons is being driven by multiple factors, including:

• Military competition and AI arms races. Between major powers (US, China, Russia), driving accelerated LAWS development with limited oversight.

• National security framing. Governments treat LAWS development as an existential security priority, subordinating ethical concerns.

• Cost efficiencies. Autonomous systems can be produced far more cheaply than manned platforms, incentivising adoption at scale.

• Speed imperatives. The pace of modern warfare creates pressure to remove slow human decision-making from targeting loops.

• Development opacity. Classified and dual-use military AI research proceeds without public accountability or international transparency.

• Regulatory vacuum. The absence of binding international law specifically governing LAWS creates permissive conditions for their development and deployment.