The Uses of Belief: A Psychoanalytic-Feminist Critique of ‘AI Ethics'

Postwar cybernetics transformed information from a measure of relation into a medium of regulation, reconfiguring communication as a problem of prediction and control. This article traces how that shift produced the modern ‘user’: a calculable figure whose rationality and adaptability were built into feedback systems linking psychology, computation, and governance. Through a close reading of Joseph Weizenbaum’s ELIZA (1966) and his reflections ‘On the Impact of the Computer on Society’ (Hamburg, 1971; Science, 1972), the paper shows how simulated dialogue substitutes plausibility for understanding, and how ethical critique becomes absorbed into the structures it seeks to address. Feminist and psychoanalytic perspectives clarify why users’ projections and affective investments sustain the appearance of machine intelligence. Contemporary AI ethics inherits these problems in translating the tension between responsibility and innovation into matters of governance and design. Recovering Weizenbaum’s ethical insight, the article argues that judgment cannot be automated without forfeiting the capacity for reflection that makes it ethical – and that avowing, rather than resolving, the contradiction between knowledge and control remains the task of critical thought in the age of AI.