More than 50 nations have joined the United States in setting guidelines to ensure the use of artificial intelligence (AI) and autonomous weapons systems adheres to international law, according to a top U.S. defense official. As AI becomes more prominent in weapons development and other military applications, stakeholders are developing rules to govern its use.
Michael C. Horowitz, U.S. deputy assistant secretary of defense for force development and emerging capabilities, provided an update on the Department of Defense’s (DOD) AI development efforts during a January 2024 virtual talk hosted by the Center for Strategic and International Studies Wadhwani Center for AI and Advanced Technologies.
In November 2023, the U.S. State Department published the Political Declaration on Responsible Military Use of Artificial Intelligence and Autonomy, which aims to establish norms of use. Fifty-one countries have endorsed the declaration as of January 2024, Horowitz said, and nations will soon begin meeting to build consensus around the use of the emerging technology, including best practices and enabling partner nations to develop their capabilities. Russia and the People’s Republic of China (PRC) have not signed the declaration, Horowitz said.
“I think there’s recognition that the sorts of norms we’re trying to promote are things that all countries should be able to get behind,” he said. “We think of this as good governance, so countries can develop and deploy AI-enabled military systems safely, which is in everybody’s interest.”
The DOD in January 2023 updated its directive regarding AI’s use in weapons systems, which calls for rigorous testing and scrutiny of all new weapons systems.
“A weapons system that isn’t safe, isn’t predictable, doesn’t work. It’s not useful,” Horowitz said. “When it comes to developing and deploying the capabilities that the joint force needs to deter war, and if necessary to prevail if conflict occurs, we need to have confidence in our systems.”
The DOD directive also calls for systems to allow human commanders and operators control over the use of force, particularly in the case of nuclear weapons, which Horowitz said adheres to the department’s Nuclear Posture Review.
“We think the decision to use nuclear weapons is so important that we think human involvement should be central,” he said. “We would expect that other countries would share that commitment and we would hope that they would make that commitment explicit.”
The DOD is working with partner nations as part of its replicator initiative, an agencywide process to accelerate delivery of innovative capabilities to warfighters. The Defense Innovation Unit (DIU) leads the initiative, announced in November 2023, with an initial focus on countering the Chinese Communist Party’s military by fielding uncrewed systems that can be deployed in large numbers and put at risk during conflict given their relatively low cost. DIU and international partners are also analyzing the use of such systems in the Russia-Ukraine war.
“We used to think about either you have precision, or you have mass. That’s no longer the case,” Horowitz said. “What we need in many instances is going to be precise mass.”