Abstract
In patent practice during the era of generative AI, human "insight " is of paramount importance. The intellectual property processes carried out by IP practitioners consist of three layers: (1) Layer 1 (invention creation and problem-solving), (2) Layer 2 (applicant and business strategy), and (3) Layer 3 (stakeholders). While AI excels at proposing new combinations of existing technologies in the first layer, it struggles with decision-making in higher layers that require context and tacit knowledge. Furthermore, insight is defined as "breadth of knowledge × depth of knowledge × experiential knowledge," and "experiential knowledge"—based on the narratives and real-world experiences of individuals or organizations as collectives of individuals—is unique to humans and cannot be replaced by AI. AI outputs are merely "non-insight, " " not-insight, " or "un-insight "; they only gain value through the insight of the human evaluator.
