Will the Advancement of AI Let Us Redefine Science?

초록

According to T. S. Kuhn, science is collective activity of solving still unsolved puzzles recognized in a specialized domain of research, and this activity is guided by one or more exemplary instances of concrete problem solving. Today we have more and more cases of scientific research in which AIs are utilized as essential tools. Though the AI agents do not have subjective motivation to solve the puzzle, they will emerge as powerful actors that solve some of the puzzles in Kuhnian way, or at least help (human-) scientists do it. This would imply that the concepts of science, scientists, and scientific research will change. Will the advancement of AI induce us to redefine science? I will discuss this question from a couple of perspectives.