US already using AI in airstrikes, will let it select weapons & targets soon, says former general

The US armed forces, especially the Air Force, are already using AI to target enemy combatants using air strikes in ongoing operations. However, they are moving to a phase where AI will not only help the US armed forces aim but also decide on the weapons package and which targets to strike, as per a report by Bloomberg.

The report quotes retired Air Force General Jack Shanahan, who believes it will take five years for the US military to fully trust and use AI engines to suggest optimal strategies and targets in battlefield scenarios.

According to General Shanahan, this timeline accounts for developing sophisticated algorithms and gradually building confidence in their deployment.

General Shanahan was the former head of Project Maven, an initiative by the US Pentagon that started in 2017 aiming to integrate AI into military operations. General Shanahan stated that currently, the emphasis is on ensuring that AI systems are only used in partially automated modes to avoid significant, catastrophic errors that could erode trust in their capabilities.

Recent developments show progress in AI integration within the military. The US Army’s 18th Airborne Corps initiated data exercises in 2020 to refine AI algorithms for target identification, leading to successful operations using various military assets like drones, fighter jets, and submarines.

Additionally, entities like the US Indo-Pacific Command, US Space Force, and allies such as the UK have also begun incorporating AI into their strategies.

Central Command, responsible for operations in the Middle East, has actively deployed AI to pinpoint targets for strikes in regions like Iraq, Syria, Yemen, and the Red Sea. Shanahan views this as a positive sign of AI normalization in military operations, with increasing recognition of its effectiveness.

Despite these advancements, the Pentagon maintains that human oversight is paramount in directing strikes, even as AI capabilities are explored across different combatant commands worldwide.

Shanahan underscores the need for ongoing experimentation and improvement to instill confidence in machine learning systems, which often require continuous refinement and oversight. He believes that as AI technologies evolve and demonstrate their value through real-world operations, trust in their use will continue to grow within the military community.