Bhaskar Mitra urges critical IR theories to redefine societal good in search

Bhaskar Mitra's latest arXiv paper, "Towards Critical IR Theories and Practices," issues a provocative challenge to the information retrieval (IR) community. Mitra argues that the field's dominant liberal frame—which focuses on neutrality, free expression, and minimal intervention—is insufficient for defining and pursuing societal good. He revives a half-century-old call from IR pioneers Belkin and Robertson, who urged researchers to develop a theoretical foundation that can determine when scientific inquiry should be restrained in the face of demands contrary to societal welfare. Mitra's central proposal is to adopt a critical frame explicitly grounded in nondomination, a concept borrowed from political philosophy that prioritizes freedom from arbitrary power. This shift would provide a clear normative basis for IR research, enabling practitioners to evaluate systems not just on efficiency or user satisfaction but on their capacity to reduce structural inequalities and power asymmetries.

The paper's timing is significant as search and recommendation systems increasingly shape public discourse, access to information, and social outcomes. By advocating for critical theory—drawing on traditions like feminism, postcolonialism, and critical race theory—Mitra aims to move IR beyond superficial fairness metrics toward a deeper engagement with power structures. The work has direct implications for how AI-driven search engines handle bias, controversial topics, and marginalized communities. If adopted, this framework could fundamentally alter how researchers measure success in IR, shifting from user engagement and click-through rates to metrics of non-domination and social equity. The paper is already generating discussion in academic circles, with implications for industry practitioners designing the next generation of AI-powered information systems.