CAN MACHINES LIE? DECEPTION, INTENTION, AND SPEAKER MEANING IN ARTIFICIAL INTELLIGENCE

Abstrak

The development of generative artificial intelligence has transformed the ways in which humans obtain information and interact through language, while simultaneously raising philosophical questions concerning truth, lying, intention, and speaker meaning in human–machine communication. This phenomenon has become increasingly relevant in social and educational contexts because AI outputs can appear convincing and resemble human conversation, even though AI cannot yet be directly assumed to possess communicative intention. This study aims to understand how users make sense of erroneous or misleading AI outputs and how they distinguish error, deception, and lying in interactions with AI. The study employs an interpretive phenomenological approach to explore users’ experiences and meaning-making in relation to AI. Data were collected through semi-structured in-depth interviews with AI users who had experience interacting with generative artificial intelligence systems and were subsequently analyzed using interpretive thematic analysis. The findings reveal four major patterns of meaning: AI errors experienced as lying, attribution of intention to machines, tension between deception and lying, and ambiguity regarding AI’s status as a speaker. The findings indicate that the experience of feeling misled does not, in itself, demonstrate that AI is lying in the philosophical sense, because lying involves questions of intention, speaker meaning, and speaker status. Theoretically, this study extends the application of Speech Act Theory to nonhuman communication, while practically strengthening the importance of AI literacy that is oriented not only toward information verification but also toward a critical understanding of agency, authority, and meaning in human–AI communication.