National Defense

US Military Used ‘Computer Vision’ To Help Pinpoint Targets To Strike In Middle East

No featured image available

The U.S. military utilized artificial intelligence to hone in on targets for its retaliatory airstrikes in the Middle East, Bloomberg reported on Monday.

U.S. Central Command (CENTCOM) forces hit over 85 targets on Feb. 2 during a series of airstrikes against Iran’s Islamic Revolutionary Guards Corps (IRGC) Quds Force and its proxy militant networks in Iraq and Syria. U.S. forces relied partly on AI algorithms that can learn on their own how to locate and identify targets, according to Schuyler Moore, CENTCOM chief technology officer, who spoke to Bloomberg.

“We’ve been using computer vision to identify where there might be threats,” Moore told Bloomberg during an interview. “We’ve certainly had more opportunities to target in the last 60 to 90 days.”

Moore noted that CENTCOM is searching for “an awful lot” of rocket launchers used by Islamic militants and terrorists in the region. She said that the AI algorithms had helped track down some of the rocket launchers used by the Iranian-backed Houthi terrorists in Yemen and in vessels in the Red Sea.

Computer vision algorithms can pinpoint and identify targets by looking at satellite imagery and other information, according to Bloomberg. The algorithms were developed under Project Maven, which was created by the Department of Defense (DOD) in 2017 to advance the integration of AI and machine learning systems in the military.

Though the DOD has confirmed that it has utilized Maven algorithms for training exercises in the past, Moore’s remarks are the clearest indication that the military is taking AI integration seriously, according to Bloomberg. CENTCOM equipped Maven in the wake of the Oct. 7 Hamas terrorist attacks against Israel that left over 1,200 civilians dead and sparked broader conflict throughout the Middle East.

“October 7th everything changed. We immediately shifted into high gear and a much higher operational tempo than we had previously,” Moore told Bloomberg, noting that it was “a pretty seamless shift” for CENTCOM to switch Maven from training to operational use.

Maven is limited in its capabilities and is being used as a way to help find targets rather than independently verify them for the military, Moore told Bloomberg. Human beings validate all of Maven’s target recommendations, knowing that the AI could make a mistake.

“It tends to be pretty obvious when something is off,” Moore told Bloomberg about Maven’s recommendations. “There is never an algorithm that’s just running, coming to a conclusion and then pushing onto the next step. Every step that involves AI has a human checking in at the end.”

The U.S. has become more deeply involved in the Middle East conflict following the death of three U.S. soldiers in a drone strike launched by Iranian-backed forces in January, an act that prompted the Feb. 2 retaliatory strikes. U.S. and allied forces are deterring waves of attacks launched by Iranian-backed forces in the region on an almost daily basis.