[Reporter’s Notebook] War Data Led to the Deaths of Elementary Students... Need for AI Ethical Standards

AI Can Also Kill Innocent Lives,
But There Is Little Discussion About How to Control and Manage It

As the sirens wailed, missiles soared into the sky with a thunderous roar. In an instant, the missiles flew off and struck Iran's missile and drone launch sites, military airfields, and other key military infrastructure. A gray building, looking like a small Lego block, quickly collapsed amid gray smoke. What looked like a scene from a movie happened in a flash. This was the footage of Tomahawk missile launches and strikes released on social media by U.S. Central Command on the 3rd (local time). Before the war with Iran, the U.S. military used artificial intelligence (AI) to select 1,000 targets. It took only 24 hours to identify the target coordinates and prioritize them. Decision-making through AI ultimately led to the death of Iran’s former Supreme Leader, Ayatollah Ali Khamenei.

The U.S. military aims for “more efficient warfare” through AI. In January, during the operation to capture Venezuelan President Nicolas Maduro, it also used Anthropic's generative AI model “Claude.” Anthropic objected to the U.S. Department of Defense using Claude, citing AI ethics, but the model was again deployed in this war with Iran. The U.S. military, on the contrary, appears to be accelerating its efforts to enhance defense capabilities through AI. On the 14th, the U.S. Army announced that it had signed a contract with defense AI company Anduril to introduce infrastructure worth up to 20 billion dollars (approximately 30 trillion won) over the next 10 years. OpenAI took the opposite stance from Anthropic, signing a contract that accepts all legal terms of use set by the U.S. Department of Defense.

There is little difference between AI and humans when it comes to killing innocent civilians. When a U.S. missile struck a girls’ elementary school in Minab, southern Iran, 175 children were killed. Donald Trump, President of the United States, tried to shift responsibility by claiming it was “an act by Iran”; even when it was revealed that missile parts found at the scene were made in the U.S., he maintained his denial. The U.S. military’s attack on the elementary school was not a result of intentional targeting or a misfire. It happened simply because the U.S. military’s geographic information was not up to date. In a situation where the accountability of AI is unclear, faulty data took precious lives.

What makes AI even more frightening is that it further trivializes our attitude toward war. Guns have made killing easier than swords, missiles more so than guns, and now AI lessens our sense of guilt in causing casualties even further. President Trump’s remark after airstrikes on Kharg Island, Iran’s main oil export hub—“I could attack a few more times just for fun”—is a prime example. We are no different. War that looks like a game has shifted the focus in our country to “stock prices” and “oil prices.” While the Middle East lives in constant fear of possible airstrikes, we are obsessed with the “numbers.”

Sam Altman, CEO of OpenAI, said on a podcast last August that “Gen Z is the luckiest generation thanks to AI,” referencing the many business opportunities AI creates. But life must always come before profit. While it has become clear that AI can also kill innocent lives, public debate about how to control and manage its use remains lacking, as all attention is focused on technological advancement. Even now, we must establish concrete ethical standards for AI and accelerate their implementation.

This content was produced with the assistance of AI translation services.