[ad_1]
Why it's important: AI on the battlefield can allow faster, more coordinated military strikes and responses. In the future, it could also be used to calculate the enemy's strategy much more accurately and efficiently. But imagine if you tried to live peacefully in a war-torn country with drones above you – could you trust computers to not go wrong?
I could not. I do not want the AI to be able to decide whether my life should be over, so I do not think that power should be given to anyone. It seems that military commanders share my opinion and refuse to trust artificial intelligence without extensive human oversight, but the Pentagon and the DARPA (Defense Advanced Research Projects Agency) have launched a new program of defense. To change that.
At a conference in Washington celebrating the 60e anniversary of DARPA, they revealed that they had allocated $ 2 billion for the development of military AI over the next five years. Although it is not much by DARPA standards, it is a lot for AI and it will lead to significant improvements. It's the most spent on AI, but it's only one of 25 simultaneous AI programs that DARPA runs. In July, they revealed that defense contractor Booz Allen Hamilton had received $ 855 million for a five-year non-AI project and that DARPA would award up to $ 1 million by research group to improve the recognition of complex environments.
The main objective is to create an artificial intelligence capable of explaining his choices to a supervisor so that he can prove that he uses common sense and has a digital "moral compass". DARPA also aims to solve problems related to the unpredictability of artificial intelligence and to prove that it can take into account unexpected variables as can a human being. Currently, the AI used can only give a "confidence rating" as a percentage of error, and it is not allowed to shoot without a human signature.
This initiative could be a successor to the Maven Project – a Pentagon initiative to improve the recognition of objects and the environment in military situations. Google was their main partner, but after an outcry from Google employees, they feared creating software that could one day be used to kill people, but Google executives decided not to renew the contract.
The Rand Corporation, another Pentagon contractor, also voiced concerns. They particularly emphasized the impact of AI on nuclear war, pointing out that the US could possibly bypass some of the foundations of the theory of mutual assured destruction. In particular, they argue that in the future, AI could be used to predict the exact location of mobile intercontinental ballistic missiles and use conventional weapons to destroy the nuclear arsenal of a foreign power. The only solution for foreign power is an immediate attack, initiating a nuclear war.
Until now, the government has not responded to these concerns. Instead, the Trump administration began building a joint artificial intelligence center to coordinate AI research across the Department of Defense.
Ron Brachman, who previously commanded DARPA's AI research, said at the conference that "we probably need a gigantic Manhattan project to create an AI system with the skills of a three-year-old child." wait and see if Trump wants to invest as much money, time and energy in the AI - or if Putin will do it first.
Source link
