Comment by soco

If we can't understand the problem, will we be able to understand the solution presented by that AI? Or we'd just apply it, trusting blindly the unfathomable reasons the AI used? Do we have an AI where the decision tree can be grasped by humans?