Why the Military Can’t Trust AI
Large Language Models Can Make Bad Decisions—and Could Trigger Nuclear War
By Max Lamparth and Jacquelyn Schneider
April 29, 2024
https://www.foreignaffairs.com/united-states/why-military-cant-trust-ai In 2022, OpenAI unveiled ChatGPT, a chatbot that uses large language models to mimic human conversations and to answer users’ questions. The chatbot’s extraordinary abilities sparked a debate about how LLMs might be used to perform other tasks—including fighting a war. Although for some, including the Global Legal Action Network, LLMs and other generative AI technologies hold the promise of more discriminate and therefore ethical uses of force, others, such as advisers from the International Committee of the Red Cross, have warned that these technologies could remove human decision-making from the most vital questions of life and death.
The U.S. Department of Defense is now seriously investigating what LLMs can do for the military. In the spring of 2022, the DOD established the Chief Digital and Artificial Intelligence Office to explore how artificial intelligence can help the armed forces. In November 2023, the Defense Department released its strategy for adopting AI technologies. It optimistically reported that “the latest advancements in data, analytics, and AI technologies enable leaders to make better decisions faster, from the boardroom to the battlefield.” Accordingly, AI-enabled technologies are now being used. U.S. troops, for example, have had AI-enabled systems select Houthi targets in the Middle East.
Both the U.S. Marine Corps and the U.S. Air Force are experimenting with LLMs, using them for war games, military planning, and basic administrative tasks. Palantir, a company that develops information technology for the DOD, has created a product that uses LLMs to manage military operations. Meanwhile, the DOD has formed a new task force to explore the use of generative AI, including LLMs, within the U.S. military.
https://www.foreignaffairs.com/united-states/why-military-cant-trust-ai
