Armed conflict in the future may look nothing like it does today. Thanks to advances in artificial intelligence, swarms of robots capable of operating at superhuman speed using AI are likely to be deployed. And as human commanders struggle to keep up with their robot combatants, the military may give AI algorithms more authority to make decisions on the use of lethal weapons. The Defense Advanced Research Projects Agency, a blue-sky research division of the Pentagon, is currently conducting experiments involving dozens of ground and aerial robots operating in simulated urban warfare scenarios. Their mission: Find terrorists suspected of hiding among several buildings. Military officials involved in the Darpa trials say the exercises prove that robots will often need to operate autonomously in order to maintain an edge over adversaries. "Is it even necessary to have a human in the loop?" US Army Futures Command general John Murray asked an audience at the US Military Academy last month. The answer is complicated. Artificial intelligence can behave unpredictably, and biased training data could perhaps cause a drone to mistake a civilian for an enemy. Failing to ban the technology could also hasten a troubling new AI-fueled arms race. But this much is clear: As AI advances, the prospect of more autonomous weapons will too. Will Knight | Senior Writer, WIRED |
Post a Comment