Today, with the dawn of Artificial Intelligence (AI), humanity has made its most potent creation yet. When the first atom was split, it opened the door to nuclear energy and nuclear medical treatment, all of which created large-scale employment. In addition, the nuclear power industry nurtured existing businesses to grow by providing a reliable, cost-effective power source and shrinking their carbon footprints, especially in energy-intensive industries. However, the splitting of the atom also led to the creation of humanity’s gravest horror: the atomic bomb. AI, a far more flexible tool, finds itself in a similar circumstance: promising tremendous economic gain but also terrifying military power.
Market.us, a market-tracking firm, predicts that the implementation of AI in the military will rise from $8.9 billion to $24.7 billion – nearly tripling in value – by 2032. This future growth is attributed to the up-and-coming sector of defense-tech start-ups, data-labeling suppliers and advanced-sensor manufacturers. Governments are encouraging AI investments for several reasons. Firstly, AI promises to augment and modernize existing military technology. For instance, Shield AI, a California start-up, says it can bring outdated F-16 fighter jets into competition with advanced stealth fighters with AI piloting. Analytics companies like Palantir have already allowed the armed forces to make more of their existing surveillance systems by using AI to sift through masses of data. Such secondary benefits also go beyond the military: the Trump administration has designated AI technology more broadly as one of the key areas of strategic competition between the U.S. and China.
Economic gain is only half the story. Colonel John Boyd, one of the most decorated U.S. fighter pilots, had a simple psychology: the OODA loop (Observe, Orient, Decide, Act). At the operational level, AI compresses Boyd’s classic strategy into milliseconds. In his 2022 paper, 5 Advantages of AI on the Modern Battlefield, Systematic Defense vice president Hans Bohlbro illustrates how AI can fuse radar, video and radio traffic to flag an incoming threat before any human assigned could glance at all of those three screens. This acceleration affects the pace and rhythm of combat. AI outstripped the primal logic of combat and elevated the cadence of decision-making involved in warfare to seconds — an innovation already visible on modern battlefields.
Militarized AI could save lives instead of taking them, too. The United States Defense Advanced Research Projects Agency’s ARES program, for instance, is building an unmanned vertical-lift machine that aims to aid in the evacuation of wounded soldiers in terrain too hazardous for manned helicopters. Additionally, machine-vision tools can spot bleeding in seconds during the battle, helping medics treat the wounded faster. Similarly, routing programs can safely guide convoys through unfamiliar terrain and territory.
However, whether the defensive and humanitarian applications of AI will outstrip control of its offensive ones remains to be seen. As AI grows to be trusted by commanders, it is not unfeasible to imagine a scenario where instead of humans, AI is alone in the boardroom. Therefore, militarized AI stands to be a double-edged blade: an effective tool both for and against humanity.
There is concern that autonomy could hollow out accountability. Human Rights Watch’s 2025 report, A Hazard to Human Rights, argues that when a computer (or a non-living being) complies with the target list, neither the coder nor the commander are legally responsible for the consequences. Recognizing this, the UN Secretary-General called in August 2024 for a binding treaty by 2026 to forbid weapons that function without human oversight. As for the more distant future, history suggests that the answer lies neither in blind embrace nor outright ban. Guardrails and guidelines are essential, especially in these early years of AI. Transparency and ethical use of AI is of paramount importance.
The AI age, like the atomic age, offers promise and peril in the same package. The challenge lies in capturing the merits of this gift and sidestepping its dangers.
Our success depends less on the brilliance of our algorithms or code than on our resolve to be morally and legally aware at the time force is applied, using AI.
