Trusting the machine: a digital humanist perspective on misplaced trust in artificial intelligence

Abstract The widespread invocation of “trust” in discussions surrounding generative AI (GenAI) conceals a deeper conceptual and normative tension. This paper critically examines the increasing tendency to treat AI systems as recipients or objects of trust, arguing that such a move risks both conceptual distortion and ethical erosion. Drawing on a digital humanist perspective, the paper contends that trust is a foundationally human, moral concept—anchored in vulnerability, autonomy, and reciprocal recognition—that cannot be meaningfully transferred to machines. The paper differentiates trust from mere reliance and explores the consequences of anthropomorphising AI systems. Using cases such as AI chatbots in therapeutic contexts, it shows how the application of trust to AI not only undermines human agency and emotional depth but also encourages a mechanisation of human social practices and an erosion of moral responsibility. In contrast to functionalist or system-oriented views of trust, a digital humanist approach insists on maintaining the conceptual boundary between humans and machines, advocating for transparency, controllability, and accountability in AI systems—without misappropriating the language of trust. The paper scrutinizes how framing AI as trustworthy reconfigures social norms, blurs lines of responsibility, and endangers the cultivation of human morality. A digital humanist reassertion of trust as an intersubjective and ethical relation is vital to resist these trends and to uphold the dignity and agency of human actors in the age of artificial intelligence.