This paper proposes four core dimensions for evaluating scientific theories: heuristic value, predictive capability, explicit constraints, and verifiability. From a philosophy-of-science perspective, we argue that the worth of a theory lies in its capacity to inspire new research directions, generate testable predictions, define its own boundaries, and remain open to empirical verification. Using historical precedents—such as Galileo’s use of the telescope and Hooke’s use of the microscope—we show that the acceptance of new tools has often met resistance, yet these tools ultimately catalyzed paradigm shifts. We extend this analogy to modern AI-assisted research, contending that AI, like the telescope or microscope, is merely an instrument; its utility depends on the underlying theoretical insight. Rejecting AI solely on the basis of its origin is philosophically equivalent to rejecting earlier scientific instruments. Finally, we identify potential biases in academic culture that may favor “purely human” work, and we argue that such preferences risk stifling innovation in the same way that historical skepticism toward new tools once did.
jiazheng liu (Wed,) studied this question.