Autonomous systems in Ukraine reveal escalating militarisation of AI, obscuring geopolitical and ethical drivers behind battlefield automation

Indigenous knowledge systems universally reject the commodification of life and the delegation of lethal decisions to machines, framing such actions as a violation of ancestral wisdom and intergenerational responsibility. The Māori concept of *kaitiakitanga* (guardianship) and the Navajo principle of *Hózhǫ́* (harmony) explicitly oppose autonomous weapons as antithetical to their cosmologies. These perspectives are systematically excluded from Western military-technological discourse, which treats Indigenous lands as testing grounds for weapons systems.

The current militarisation of AI mirrors historical arms races, from the development of gunpowder to nuclear weapons, where technological superiority was conflated with strategic dominance. The 19th-century industrialisation of warfare (e.g., machine guns, tanks) similarly obscured the human cost behind claims of 'efficiency' and 'precision.' The lack of international treaties on autonomous weapons today echoes the failure to regulate chemical weapons in the interwar period, suggesting a pattern of delayed governance until catastrophic use cases emerge.

Cross-culturally, the ethical rejection of autonomous weapons spans Confucian traditions (which prioritise human judgment in moral dilemmas) and Islamic jurisprudence (which requires *taklif*, or moral accountability, in warfare). In Latin America, the *Tupamaros* and other revolutionary movements historically rejected indiscriminate violence, a principle that could inform global norms on AI in warfare. These diverse ethical frameworks highlight a shared concern: the erosion of human agency in life-and-death decisions.

Scientific consensus warns that autonomous weapons lack the contextual understanding to comply with international humanitarian law, particularly in distinguishing combatants from civilians or assessing proportionality. Studies by the Future of Life Institute and IEEE show that current AI systems exhibit bias in target recognition, with error rates disproportionately affecting marginalised groups. The 'Terminator' narrative distracts from these technical limitations, which are far more immediate threats than hypothetical superintelligence.

Artistic and spiritual traditions have long warned against the dehumanisation inherent in mechanised warfare, from Picasso’s *Guernica* to the poetry of Wilfred Owen. In Buddhist thought, the cycle of violence (*samsara*) is perpetuated by attachment to weapons, while in Christian theology, the just war tradition requires human judgment (*discernment*) that machines cannot replicate. These traditions frame autonomous weapons as a spiritual crisis, not merely a technological one.

Future scenarios predict that unchecked militarisation of AI could lead to 'flash wars' where autonomous systems escalate conflicts beyond human control, akin to the 1983 Soviet nuclear false alarm incident. Modelling by the RAND Corporation suggests that even 'defensive' autonomous systems could trigger preemptive strikes due to misinterpreted signals. The long-term risk is not Skynet, but a world where warfare becomes a self-perpetuating algorithmic process, divorced from political accountability.

Marginalised communities in conflict zones—particularly women, children, and Indigenous peoples—are the first to suffer from autonomous weapons, yet their testimonies are excluded from Western media narratives. In Yemen, drone strikes have disproportionately targeted civilian areas, with little recourse for survivors. The framing of AI as 'supportive' rather than 'lethal' further silences those who experience its consequences, reinforcing a hierarchy where technologically advanced nations dictate the terms of warfare.