Reading without knowing: how authorship disclosure shapes ethical engagement with AI-translated literature

As generative artificial intelligence (GenAI) increasingly produces literary texts, ethical questions arise about how readers emotionally engage with AI-generated content, particularly when authorship is uncertain. This mixed-method study compares readers’ empathy toward human- and AI-translated literary excerpts. Quantitative results showed no significant differences in affective, cognitive or associative empathy, or in perceived quality, between conditions. Qualitative analysis revealed that readers’ engagement was driven primarily by textual features rather than authorship assumptions. However, when prompted to consider possible AI involvement, some participants reconsidered their judgments through attributional reasoning about machine capabilities. The findings suggest that authorship disclosure shapes interpretive autonomy and evaluative judgments of machine-generated texts, with important ethical implications.