(Repositioning) AI as trickster in organizations for bias mitigation

Abstract

As a contribution to the ongoing discussions on artificial intelligence (AI) ethics, I propose a feminist way of thinking on the positioning of AI in organizations. Inspired by the trickster figure that originates in mythology and folklore, I develop the notion of ‘trickster AI’ to conceptualize the fluid, relational, and ambivalent role of AI in organizational contexts. The ‘trickster AI’ is further supported by an empirical study drawing on a series of co-creation workshops on AI design for bias mitigation in hiring. Four AI-based solutions from the workshop – neutralization , standardization , customization , and diversification – are identified and analysed to examine how different configurations of AI and human actors shape the construction of bias and fairness. Rather than treating AI as inherently fair or biased, the findings show how the role of AI is continuously negotiated and reconfigured across practices. Interpreting these configurations through the lens of the trickster, the article highlights AI’s fluid positionality, inherent ambivalence, and capacity for boundary crossing. It contributes to feminist organization studies and feminist approaches to AI by moving beyond binary understandings of technology and foregrounding the situated, relational, and multiple enactments of fairness in sociotechnical practices.