Greek Mythology as a Lens for Understanding the Risks, Biases, and Ethics of Artificial Intelligence

Abstract

As artificial intelligence systems become deeply embedded in everyday life, their risks increasingly extend beyond technical failures or localized biases to encompass broader sociotechnical and structural dynamics. These dynamics-such as the destabilization of trust, the amplification of power asymmetries, and the erosion of human agency-are often difficult to communicate to non-specialist audiences without resorting either to technocratic abstraction or to alarmist rhetoric. This paper explores Greek mythology as an interpretive and pedagogical lens for making the risks, biases, and ethical challenges of artificial intelligence more intelligible and critically accessible. Drawing on the mythological corpus of ancient Greece, we argue that myths constitute widely shared narrative frameworks that can be mobilized to reflect on responsibility, the exercise of power, decision-making, and their collective consequences. By engaging with selected myths-most notably the figure and fate of Oedipus-we analyze how their narrative structures can illuminate contemporary AI-related issues, including inappropriate reliance and overtrust in computational systems, opaque decision-making processes, feedback loops that reinforce inequality, and the moral risks associated with delegated authority. Treating mythology not as a mere metaphor but as a shared symbolic language that bridges technical expertise and insights from the social sciences and humanities, this paper situates mythological narratives within a sociotechnical perspective. It shows how such narratives can function as tools for critical foresight and ethical reflection across diverse audiences, ranging from young users to policymakers and regulators. We argue that these narrative frameworks can foster more reflective, inclusive, and context-sensitive approaches to AI evaluation and governance in a context marked by rapid technological transformation and growing societal dependence on computational systems.