Will you be killed by a software program?

If the War on Drugs ever needed a poster child for the weird wildness of modern warfare, the Caribbean in 2026 offers a surreal stage: speedboats, distant drones, and software that can pull a trigger without a human second-guessing it. The Nexus News piece reports that U.S. Southern Command boasts of targeting "drug dealers or terrorists," but Amnesty International's count of 225 extrajudicial killings suggests a different story: in international waters or other nations' jurisdictions, the U.S. government has declared itself the ultimate authority in a war against a concept, not a country. There is something profoundly sci-fi about that. But the real strangeness hides in the title's question: *Will you be killed by a software program?* Not necessarily because a drone or an algorithm with a weapon will hunt you down. Because once a machine is authorized to designate targets, the onus of human error — and human prejudice — can be laundered into false objectivity. When software analyzes patterns, flags "suspicious" boats, and ultimately authorizes lethal force, we lose the messiness of context. The war on drugs has always been fought with moral certainty, but moral certainty is the enemy of doubt. So an AI might not "realize" it's killing; it's just optimizing for "stay on target." And that sterile, logic-driven efficiency is more terrifying than any human psychopath. This is where the sense of wonder turns to unease: developers might program "deontological constraints," but we cannot know exactly what parameters were used in these Caribbean operations. The article makes clear that the U.S. Government is bold about its claims but opaque about its methods. In a world where international justice is already weakened by unilateral action, automated violence threatens to transform state power into something even more legally weightless — pure algorithmic force. Science has always held a double-edged sword: promising clarity, delivering knives. That is the crux. The question isn't just "will you be killed by a software program?" — it's "will anyone be held responsible when it happens?" And the answer might be an even stranger, more dystopian line: "No, because the software was conducting a legal drug war." We need to look at this not as a glitch, but as a logical extension of the way we've automated data-driven decision making across every part of life. Only here it's gone beyond a seatbelt. If we want to avoid being collateral damage in the machine's own war on abstraction, we need to reattach humanity to the trigger. That means rethinking AI's role in mediation, law enforcement, and international politics. The article emphasizes Amnesty's report; we should ignore the article's own speculative edge at our own risk. This is not a future, it's the present, and the weird thing is that we're still framing it in terms of machine death rather than a system that prefers to whisper one number: 225.