Without a start in the face of AI, “major incidents” could occur within one to two years, warns Google DeepMind

When you run one of the leading AI labs on the planet, words aren't chosen at random. Artificial intelligence agents could cause “major incidents” within one to two years, if companies and states do not do what is necessary to make them safe, Demis Hassabis, president of Google DeepMind, warned on Wednesday.

The loss of control of autonomous agents observed in recent months “are in a way early warning signals, because these systems are not yet very powerful, so the incidents are not major, but they could be major within one to two years if we do not increase efforts” in terms of security, he declared during an AI summit in Dublin.