Профиль

This motivated me last year to start developing a research agenda for emancipatory IR. It is my hope that universal humanization and emancipation from oppression can be a kernel that we can build a broader coalition within the IR community and mobilize towards. Paper: irrj.org/article/view... 50/

#3 IR needs to make a normative commitment to nondomination in solidarity with the oppressed. This is a challenging task for a community whose members reflect vastly different values, cultures, politics, identities, privileges, lived experiences, and understandings of history. 49/

IR should reconnect with its roots in library and information sciences that engages w/ liberatory concerns. Libraries have been at the forefront of antifascist and decolonial resistance throughout history across the globe. (Shout out to @patchedelbows.bsky.social for inviting me to #CALC26!) 48/

To do this work appropriately, IR needs interdisciplinary collaborations, to engage with critical / feminist / decolonial / other liberatory scholarships, and be informed by the body of knowledge from activism and movement organizing. 47/

We also need to consider the systemic consequences of emerging tech like AI for information access. While concerns of hallucinations and bias are important, we mustn't overlook the other mechanisms by which AI disrupts the information ecosystem, concentrates power, and marginalizes communities. 46/

Another example involves the temptation to address the ecological cost of large AI models by building more efficient models without acknowledging that this is likely to intensify (not decrease) overall resource usage and ecological impact of AI due to well-known phenomenon of Jevons paradox. 45/

Eg, there has been great progress recently on aligning LLMs to "positive societal values" without appropriately addressing who gets to choose these "universal" values. Without an enforceable framework for democratically selecting the values this progress is easily coopted for mass manipulation. 44/

#2 For IR research to be rigorous, it must be informed by a systemic understanding of our world, incl. how systemic power relations manifests (and is maintained) between groups and systems of oppression. Otherwise we end up with highly counterproductive outcomes despite positive intentions. 43/

#1 We need to reject technological determinism / AI-inevitability narratives and take back our agency to shape technology for societal good. We must also critically interrogate and challenge the dominant + oppressive power structures that shape our visions of our desired sociotechnical futures. 42/

I make 6 recommendations for Critical IR in this context based on couple of my recent / ongoing work: **Towards Critical IR Theories and Practices** arxiv.org/abs/2606.30984 **Information Access of the Oppressed: Freirean Design for Emancipatory Information Access** arxiv.org/abs/2601.09600 41/

We see this shift from liberal to critical frame in IR—"FACT IR: Fairness, Accountability, Confidentiality and Transparency in Information Retrieval" (SWIRL'18) → "Centering Societal, Democratic, and Emancipatory Values and Ethics in IR" (SWIRL'25), and workshops like jedi.inertial.science. 40/

...is explicitly political and politically decolonial, anti-racist, feminist, queer, pro-labor, anti-capitalist, anti-casteist, anti-ableist, and abolitionist. It is philosophically social constructionist and principally nondominative. See article: arxiv.org/abs/2606.30984 39/

We find 2 distinct frames in IR today: the dominant liberal frame and the emerging critical frame. The former wants good systems, focuses on harm mitigation, and tries to appear apolitical. The latter wants good society (just/emancipated/democratic), leaves room for repair & rebuilding, and... 38/

Similarly, with respect to values, politics, and epistemologies, there is tension between explicating them vs. keeping them implicit, from projecting an apolitical view of our research vs. recognizing its saliently political nature. 37/

With respect to methods, are we limited to trying to reduce the harm of emerging systems *or* are we trying to design societally beneficial systems? The risk of the former is that we get stuck at an equilibrium where harmful systems are rendered palatable _enough_ through ethics-washing. 36/

I think the issue of political bias in search is not a fairness problem. The underlying problem is the platform's concentration of power on deciding what ideas deserve exposure and the total lack of space for social deliberation and negotiation within these systems to inform these choices. 35/

Without critical perspectives, we may inappropriately apply fairness constructs—e.g., papers that uncritically promote algorithmic bothsidism as I argue in irrj.org/article/view.... IR should NOT give equal exposure to democratic & fascist values, nor to gender justice & misogyny/transphobia. 34/

There is a tension between trying to set our goals with respect to what constitutes a *good system* (e.g., fair and transparent) vs. a *good society* (e.g., just and democratic). I argue we cannot meaningfully define the former (nor validate our constructs impact) without defining the latter. 33/

To theorize the concept of "societal good" we need to critically interrogate the frames through which IR operates. I loosely define a frame as comprising of a set of desired goals, a collection of methods and practices, and a body of underlying values, politics, and epistemologies. 32/

Last week I presented a keynote at #SIGIR2026 titled "Emancipatory Information Retrieval: Towards Critical IR Theories and Practices". Slides: bhaskar-mitra.github.io/files/emanci... I went in nervously excited but several people appreciated talking openly about this topic. A quick thread. 🧵 1/