A defense official informed Bloomberg News that the US utilized artificial intelligence to pinpoint targets struck by air strikes in the Middle East this month.
They were showcasing an increasing military application of this advancing technology in warfare.
Schuyler Moore, the chief technology officer for US Central Command, stated that machine-learning algorithms capable of self-teaching were used to pinpoint targets for over 85 US air strikes on February 2 in the Middle East.
The Pentagon stated that bombers and fighter planes carried out the strikes on seven locations in Iraq and Syria in response to a lethal attack on US service members at a Jordanian base.
Ms. Moore informed Bloomberg News that they have been utilizing computer vision to detect potential threats.
We have had increased opportunities for targeting in the past 60 to 90 days.
She stated that the US is seeking a significant number of rocket launchers from hostile groups in the region.
Before this, the military has admitted to utilizing computer-vision algorithms for intelligence objectives. Yet, Ms. Moore’s statements provide the most definitive proof of the US military employing this technology to pinpoint enemy targets that were later attacked.
The US airstrikes, as confirmed by the Pentagon, targeted and either destroyed or damaged rockets, missiles, drone storage facilities, and militia operations centers. These attacks were carried out in retaliation to the killing of three service personnel in an attack on a military base in Jordan on January 28, as part of President Joe Biden’s actions.
The US credited the incident to Iranian-supported militants.
Ms. Moore mentioned that AI technologies have assisted in detecting rocket launchers in Yemen and surface vessels in the Red Sea. Centcom reported destroying some of these targets in weapons attacks this month.
Iran-backed Houthi forces in Yemen have launched many rocket attacks on commercial vessels in the Red Sea.
Targeting algorithms were created as part of Project Maven, a Pentagon program launched in 2017 to speed up AI and machine learning use in the Defense Department, particularly in supporting defense intelligence and developing prototypes for the US efforts against ISIS militants.
Ms. Moore stated that US military personnel in the Middle East had tested computer-vision algorithms capable of identifying and locating targets using satellite images and other data sources during exercises conducted over the last year.
Ms. Moore stated that everything changed on October 7 due to the Hamas strike on Israel that occurred before the Gaza conflict.
We quickly increased our pace and operational intensity significantly compared to before.
After a year of digital training, US soldiers could transition to employing Maven.
Ms. Moore highlighted that Maven’s AI capabilities are utilized to identify prospective targets, not verify them or deploy munitions.
Exercises conducted by Centcom late last year involved testing an AI recommendation engine, which demonstrated that these systems often did not perform as well as people in suggesting the sequence of attack or the most effective weapon to utilize.
Humans review the AI-targeted recommendations, she stated. US operators are diligent in acknowledging their duties and the potential for errors with AI since they can detect when something is not functioning well.
She stated that algorithms do not run, reach a decision, and move on to the next stage.
At the final stage, a human oversees every AI process.
