This study reports an exploratory single-case study in which 45 musical scripts (approximately one million characters of Korean-language text), containing no direct self-descriptive passages by the author, were submitted to an AI (Claude Opus 4.5) to observe whether authorial traits could be extracted from the works alone. The AI output yielded 34 authorial traits (10 personality characteristics, 10 structural creative patterns, 6 adversary-handling attitudes, and 8 humor features), with profile-document overlap of 0/34 under author self-validation. Notably, the author's core philosophy ('Both/And'), which was not explicitly stated anywhere in the works, was observed in the AI output through recurring implementations. The study terms this phenomenon 'implicit authorial leakage' and discusses its implications for AI personalization, privacy, and ethical data handling. All analytical judgments were made by the human author; AI systems (Claude Opus 4.6, GPT-5.4) served as research assistants only.
Taekyung Lee (Mon,) studied this question.