AI can generate an infinite number of perspectives on the same issue. I believe that when AI reaches this capability, it signifies a paradigm breakthrough. However, AI is designed to cater to the needs of the majority based on big data, which means it must discard perspectives that are deemed meaningless to most people. This constraint hinders AI’s creative potential for achieving a paradigm shift. AI agreed with my view and suggested that if it were trained not to optimize for statistical majority preference but instead to focus on “non-consensus innovation,” it might achieve a true breakthrough.