As the US military accelerates its adoption of autonomous weapons systems amid a growing global artificial intelligence arms race, one expert told Common Dreams on Wednesday that \u201cgreater action needs to be taken urgently\u201d to protect civilians and ensure meaningful human control over rapidly developing technologies.<\/p>\n
US Defense Secretary Pete Hegseth told congressional lawmakers Wednesday during a House Armed Services Committee hearing <\/a>on the proposed<\/a> $1.5 trillion Pentagon budget for 2027 that the military will soon have a new \u201csub-unified command\u201d dedicated to autonomous warfare.<\/p>\n Hegseth, who advocates \u201cmaximum lethality<\/a>\u201d for US forces, has expressed disdain<\/a> for what he called \u201cstupid rules of engagement\u201d designed to minimize civilian harm. He has overseen the dismantling of efforts meant to mitigate wartime harm to civilians \u2013 hundreds of thousands of whom have been killed in US-led wars during this century, according to<\/a> experts.<\/p>\n This \u201cmaximum lethality\u201d ethos, combined with AI-powered systems allowing for exponentially faster and more numerous target selection, has raised concerns that have been underscored by actions including Israel Defense Forces massacres in Gaza and Lebanon, and US attacks like the cruise missile strike<\/a> on a school in Iran that killed 155 children and staff.<\/p>\n \u201cA sole focus on achieving maximum lethality is inherently incompatible with civilian protection,\u201d Verity Coyle, deputy director of Human Rights Watch\u2019s (HRW) crisis, conflict, and arms division, told Common Dreams. \u201cIf the United States truly seeks to protect civilians, it should forgo this limited focus and ensure it has guardrails in place that assess the proportionality of its actions and guarantee a distinction between civilians and combatants.\u201d<\/p>\n \u201cUnder international humanitarian law, civilian protection requires that military actions abide by the principles of distinction and proportionality,\u201d Coyle noted. \u201cIn other words, military actors must distinguish between civilians and combatants and ensure that the resulting harm to civilians from their actions would not be excessive in comparison to the perceived military gain.\u201d<\/p>\n Experts on lethal autonomous weapons systems \u2013 commonly called \u201ckiller robots\u201d \u2013 stress the need for meaningful human control. However, with industry-backed efforts afoot<\/a> to ban state and local governments from placing guardrails on AI development, retaining such control could become increasingly difficult as the technology advances.<\/p>\n \u201cThe lack of serious guardrails… shows a troubling lack of concern for these real and immediate risks to civilians both in the United States and abroad,\u201d Coyle said. \u201cWhile we have seen some Congress members and state legislators express concern over these developments, greater action needs to be taken urgently.\u201d<\/p>\n Asked about the \u201cif we don\u2019t build it, they will\u201d mentality of many US proponents of unchecked AI development that is reminiscent of the Cold War nuclear arms race, Coyle said the United States is ignoring its \u201cability to set the global agenda and international humanitarian law norms.\u201d<\/p>\n \u201cAs we see greater integration of AI in the military domain and resulting civilian harm, we need strong international leadership to respond to these threats, not states relinquishing their responsibilities,\u201d she asserted.<\/p>\n Coyle continued:<\/p>\n Throughout [HRW\u2019s] decades of work in banning weapons that cause indiscriminate civilian harm, including the Mine Ban Treaty and Convention on Cluster Munitions, we have seen that even when some major military powers object to new international law, other states are able to band together and create new norms that major military powers eventually abide by. In this moment, the United States needs to decide if it will stand up for the principles of civilian protection and a rules-based order, or if it will walk away from the system it helped create and that has served to protect civilians for several decades.<\/p><\/blockquote>\n There is also a danger that companies will proceed with risky AI weapons development, both in pursuit of profit and out of fear of getting left behind if they don\u2019t push forward. For example, Anthropic \u2013 maker of the AI assistant Claude \u2013 lost a $200 million Pentagon contract and is facing a government blacklist and legal battles after the company refused<\/a> to loosen safety restrictions on autonomous weapons and surveillance.<\/p>\n Meanwhile, OpenAI, which makes the generative AI platform ChatGPT, rewrote<\/a> its \u201cno military use\u201d policy to allow \u201cnational security\u201d applications of its products, opening the door to lucrative Pentagon contracts.<\/p>\n Asked what civil society can do now to rein in reckless AI development, Coyle said that while HRW remains \u201cfocused on educating decision-makers and the public,\u201d there are \u201cclear steps states can take, including supporting an international legally binding instrument on autonomous weapons systems and regulating the military use of AI.\u201d<\/p>\n