Distinguishing Human vs. AI-Generated Texts: How Humour and Emotional Expression Shape Perceived Authorship

The rapid advancement of artificial intelligence (AI) has blurred the boundaries between human- and machine-generated content, raising critical questions about trust, authenticity, and the ability to distinguish AI outputs from human authorship. This study examines how textual humor and emotional expressions influence perceptions of authorship, specifically the ability to differentiate between human- and AI-generated texts. A total of 212 participants participated in an online experiment, evaluating a series of text messages varying across three dimensions: humor (present/absent), emotional expression (positive, negative, neutral), and authorship (human/AI). Using Likert scales, participants rated the likelihood of human authorship and provided open-ended responses highlighting the factors guiding their judgments. A within-subjects design featuring 12 experimental conditions enabled robust comparisons while controlling for individual variability. Findings aim to reveal whether humor and emotional tone enhance perceptions of human authorship, offering valuable insights into human-AI interactions and the role of emotional content in shaping authenticity judgments.