The use of AI in the military context is becoming more and more common. This is shown by the latest operations launched by the United States in the Middle East.
Despite public fears, it was pretty clear that artificial intelligence technology was at some point going to become a tool for military actions. In the United States, the Pentagon has launched several initiatives to benefit from the contributions of AI. And we are starting to see the results, having been used recently in offensive actions by the American army.
85 strikes carried out using AI
Artificial intelligence has not been idle in the premises of the United States Department of Defense. Because according to comments from the head of technology within the United States Central Command, Schuyler Moore, comments reported by Bloomberg Newsmany strikes carried out by the USA have recently been guided by AI.
More precisely, 85 strikes on sites in the Middle East during the month of February are concerned. In this context, AI helped American soldiers distinguish between centers where rockets, missiles, drones and militia operations centers were located during attacks launched on February 2 in Iraq and Syria. The technology is also used to strike the Houthi rebels, who, in support of the Gazans, have for several months been the source of fire against ships in the Red Sea.
A former project in collaboration with Google
For these needs, the Pentagon used machine learning algorithms that had been developed as part of the Maven project. A project which, at its birth, benefited from a partnership with Google, before being the subject of harsh criticism and a petition gathering several thousand signatures from employees of the Californian firm.
Management thus had to abandon the collaboration with the Pentagon at the end of the contract, in 2019. The end of an association which did not prevent the American military, again according to Schuyler Moore, from continuing to explore the best way to use AI to spot targets, using images produced by drones and satellites.
Schuyler Moore still wants to point out that we are still far from Skynet, and that the AI is constantly supervised by humans, humans who choose how to carry out the operation and what weapons to use. “ Each AI step is controlled by a human at the end » she says. But until when ?
Source : Engadget
4