The Pentagon Will Use Artificial Intelligence To Find Nuclear Missiles And Predict Potential Strikes

Jun 7, 2018
Originally published on June 7, 2018 2:00 pm

From Texas Standard.

Reuters recently reported that the U.S. may be engaged in a new kind of arms race against countries like Russia and China, to develop an artificial intelligence program that could defend the U.S. against potential nuclear missile strikes, by anticipating where or when they are likely to occur.

Phil Stewart, the military and intelligence correspondent for Reuters, says the new AI projects are being developed to search through data streams to find the locations of mobile missiles. One prototype that has had some successes.

“It’s gone from sort of an idea – from a concept – to an early prototype,” he says, “But there will be big advances in the coming years that could eventually lead to people being confident enough in the technology to actually use it in the war room.”

In theory, the AI program could allow the Pentagon to detect where missiles belonging to hostile countries are currently located and predict their future locations. Stewart compares the AI project to long-term weather forecasting, or scans that can detect tumors before they grow.

“The idea here is to use this technology to sift through data and see patterns that would escape a human analyst,” Stewart says.

Stewart reports that the Pentagon was hiding the details about the AI program under a “layer of near impenetrable jargon” in their latest budget. Stewart says because the Pentagon has to fund these projects, they have to include some details about them when requesting money from Congress.

“In one of the projects they actually mentioned using algorithms to help with the rogue mobile missile threat,” he says.

Parts of the new AI program could be in effect by the early 2020s. However, there are some concerns the artificial intelligence-powered programs could make mistakes that could potentially harm innocents.

“The idea here was that when this thing becomes viable – and you can debate when that’s gonna be – you wanna make sure that an AI alert doesn’t automatically put U.S. missile forces on alert,” he says, “There’s a humans somewhere in there to say, ‘yep, that’s real.’”

Written by Amber Chavez.

Copyright 2018 KUT 90.5. To see more, visit KUT 90.5.