The US military has increased their spending on AI technologies capable of finding and tracking mobile and hidden nuclear missile launchers, as well as capable of detecting preparations for a nuclear strike, Reuters reported, citing several anonymous sources working on the projects. According to the media outlet, there are numerous classified projects that are "buried under a layer of near impenetrable jargon in the latest Pentagon budget."
The news outlet's sources claim that in case of the success of these projects, AI will be able to track the movements of mobile launchers and warn the military command in advance about upcoming launches of nuclear missiles, giving them time to come up with a solution to the problem. At least one of the AI-based programs has received $83 million in funding for the next year, which is relatively small, according to Reuters.
READ MORE: US Navy Concerned Russia, China Could Trick Spy Satellites, Drones — Reports
One anonymous source revealed that at least one project is aimed at tracking DPRK nuclear missiles. Another anonymous source noted that one AI project is currently undergoing testing by the US military.
Experts in AI and public policy, such as the Rand Corporation, warn about possible drawbacks of pursuing the implementation of AI technology in the nuclear military sphere. Some of them forecast the start of a new arms race between the US, Russia and also China. Others simply note that AI can possibly be tricked, as happened during an experiment conducted by MIT students, who deceived an advanced Google image classifier into believing that a plastic turtle was actually a rifle.
READ MORE: White House Silent as China Details Plans to Wrest AI Dominance from the US
In May, information surfaced that the US Navy is studying the capabilities of major states, such as Russia or China, to conceal their weapons, including missiles and tanks, from US satellites. For example, the US Navy suggested that a tank could be camouflaged to look like a car and vice versa.