For those who don’t know, the U.S. is one of the countries that have declared that they will always keep control of nuclear weapons in the hands of humans, not AI, for the peace of mind of millions of people.
Subscribe to the Softonic newsletter and get the latest in tech, gaming, entertainment and deals right in your inbox.
Subscribe (it's FREE) ►But now the Pentagon is not reluctant to use artificial intelligence to “enhance” nuclear command, control, and communication systems, which is concerning, as the line between human control and machine control is very thin.
At the end of last month, the head of the U.S. Strategic Command, Air Force General Anthony J. Cotton, stated that the command was “exploring all possible technologies, techniques, and methods to contribute to the modernization of our NC3 capabilities.”
The future seems to belong to AI, even in weaponry
In recent years, several AI-controlled weapons systems and military vehicles have been developed, such as fighter jets, drones, and machine guns. However, their use on the battlefield raises concerns, so the prospect of AI, which still makes many mistakes, being part of a nuclear weapons system seems like a nightmare scenario from Hollywood science fiction.
Cotton tried to alleviate those fears at the 2024 Conference of the Department of Defense Intelligence Information System. He said (via Air & Space Forces Magazine) that although AI will enhance nuclear command and control decision-making capabilities, “we must never allow artificial intelligence to make those decisions for us.”
Already in May, Paul Dean, an arms control official at the State Department, stated in a briefing that Washington has made a “clear and firm commitment” to keep humans in control of nuclear weapons.
Dean added that both Great Britain and France have made the same commitment. Dean said that the U.S. would welcome a similar statement from China and the Russian Federation.
Cotton stated that the increase in threats, the avalanche of sensor data, and cybersecurity issues make the use of AI a necessity to keep U.S. forces ahead of those who intend to challenge the United States.
“Advanced systems can inform us more quickly and efficiently,” he said, emphasizing once again that “we must always keep a human decision in the loop to maximize the adoption of these capabilities and maintain our advantage over our adversaries.” Cotton also spoke about using AI to give leaders more “decision space.”
For the Pentagon, the use of AI could help quickly match, interpret, and present all the data collected by these systems.
Although the nuclear launch codes are not figuratively handed over, the use of AI in any nuclear weapons system could be risky, something that, according to Cotton, must be addressed.
In February, the researchers conducted simulations of international conflicts with five different LLMs: GPT-4, GPT 3.5, Claude 2.0, Llama-2-Chat, and GPT-4-Base and found that the systems often escalated the war and, in several cases, deployed nuclear weapons without prior warning.
