“Yoshua Bengio asserts that AI safety is a global issue because a dangerous model, such as Mythos, could be designed in one country and then used by another country to attack a third.”