top of page

Contesting Colonial Technologies: Historical Models for Resistance

As Amy Woodson-Boulton observes in her introduction to this series, AI’s ability to categorize and repurpose is the product of colonial and imperial legacies. The labor of tagging and categorization that facilitates generative AI relies on exploitative economies in the Global South. These economies enact tangible, lasting harms on individuals and communities. But the urge to categorize likewise taps into colonial logics that seek to constrain and order people and places deemed unruly. Through imagined or purported cartographies, the construction of elaborate racial hierarchies, the development of bureaucracy and surveillance in the form of the information state, and more, empires wrought and continue to work the illusion of order and universal knowledge. As Anushmita Mohanty demonstrates in her piece, the modern humanities classroom has inherited these logics and superficially rehabilitated them as “scholarly” and “professional” expectations. But then and now, the ability to constrain is truly an illusion, masking what Lauren Benton describes as “variegated spaces with an uncertain relation to imperial power.”

 

If generative AI’s tendency to stereotype and reduce mimics colonial and imperial logics, then perhaps the opposite holds true as well. Do the strategies employed by subjects of empire offer productive models of resistance and subversion? In this piece, I offer a few vignettes from early modern Ireland that suggest ways we might respond, critique, and contest AI infrastructures and ideologies.


An image shows a two-page spread of colorful red and yellow line drawings. On the left page, an image of St. Matthew stylized in the style of an early 9th century illuminated manuscript is shown, with a decorative pattern framing the page. On the right side, an illustration of a lion, mid-jump, is featured in the middle of a page with a simple red box pattern on the page behind it.
Symbols of Saints Matthew and Mark (9th century) from Fac-similes of the Miniatures & Ornaments of Anglo-saxon & Irish Manuscripts by John O. Westwood, 1868. Image via Public Domain Archive.

 

Resisting Categorization

1583: Sir Owen MacCarthy Reagh writes to Queen Elizabeth I and requests reimbursement for his significant expenditures incurred in helping to suppress the bloody Second Desmond Rebellion. To Elizabeth, he presents himself as an aristocrat, a loyal adherent, a humble servant, employing the conventions of the Tudor court.

 

1584: Acting not as Tudor lord but as Irish king, Owen (or more appropriately Eóghan) inaugurates the O’Donovan, bestowing the white rod of lordship on his subordinate – a ceremony revived or constructed from a pre-Norman past, which perpetuates a hierarchy that circumvents the English crown entirely.

 

England endeavored repeatedly to order, categorize, and sometimes segregate Ireland’s diverse population. In the 12th century, Gerald of Wales’ post-conquest propaganda construed Ireland – both people and land – as barbarous, inferior, and in need of civilization and order. In the 14th century, legislation like the Statutes of Kilkenny attempted unsuccessfully to erect firm ethnic categories and boundaries. In the 16th century, bureaucratic practices like surrender and regrant sought to change the legal foundations of property and title, inviting Irish lords to surrender their lands to the crown and have them granted back under English law rather than Irish. This is a well-established element of the colonial playbook. We might look, for instance, to las castas, the elaborate racial hierarchy constructed by Spanish imperialists and to the blood quantum laws by which the United States sought to constrain membership in indigenous nations.

 

Today, we see AI replicate and amplify stereotypes and racist tropes rooted in those same false categories. Even when algorithms are adjusted to combat stereotypes through what Gilliard terms “synthetic inclusion,” generative AI tools produce appallingly racist outputs. The results of such synthetic inclusion, which seek to constrain prompts and outputs without addressing the underlying biases in the data, produce faux-diversity, as in the case of racially diverse Nazis. But examples like Owen offer us another model. His flexible self-representation resists firm categorization. He wears two identities (in fact, like all of us, myriad!), which from the perspective of empire seem incompatible, rooted in mutually opposed systems of authority. But for Owen, there is no inherent contradiction. His malleable identity offers an implicit rebuke of rigid categories and of imperial efforts to separate, suppress, and constrain. Looking to historical examples of resistance to categorization offers a far more meaningful path to diverse representation than algorithmic constraint that manifest bias in horrifyingly innovative ways.

 

The limitations of Owen’s example are also instructive. His ability to safely project those apparently competing identities, and to be taken seriously, were the product of his political, social, and economic status. The lord of a large and resource-rich barony, he possessed capital (literal and metaphoric) that he leveraged to cushion his boundary-crossing. To whom do we permit such flexibilities today? How is transgression of socially constructed categories weaponized? 

 

Seeking Authority and Authenticity

1632: Four scholars, led by Franciscan lay brother Mícheál Ó Cléirigh, embark on an ambitious project to compile an authoritative and comprehensive history of Ireland, synthesizing annals and other historical texts into a single composite known as the Annals of the Kingdom of Ireland.

 

1634: Geoffrey Keating, a priest and scholar, completes his Foras Feasa ar Éirinn. While the bulk of the work comprises a history of pre-Conquest Ireland, the preface takes aim at Keating’s contemporaries who present deliberately and inaccurately derogatory narratives of Ireland’s past and people. His narrative, he argues, provides a necessary corrective.

 

These works provide powerful exempla for us as scholars as we respond to the ways generative AI devalues knowledge, expertise, and lived experience. That devaluation is embedded in the very logic of the tool, which strives for plausibility without regard for truth or meaning. However, the devaluation manifests in other, more nefarious ways. In mimicking the vocabulary and rhetoric of workers in the Global South, generative AI appropriates their voices and thus harms their reputation. By training models on widely available content, AI companies amplify existing linguistic inequities, exacerbating issues of representation and access in the digital landscape. Bots scraping the web overtax the fragile digital infrastructures of cultural heritage institutions and inhibit access to archives.

 

Confronted with inaccurate, biased narratives that serve unjust power structures, both Keating and the Four Masters return directly to the primary sources and construct narratives that speak to their identities, their communities, their values. They work to challenge existing hierarchies by seeking alternate sources of authority.  In Keating, there is an explicit recognition that history done uncritically or maliciously can harm. In the Four Masters, there is a deep commitment to authenticity and accuracy. Both share the conviction that history is not neutral.

 

Small, Persistent Acts of Care

1595: At Timoleague Friary, Fr. Maurice O’Hea is summoned by an invalid who has given up all hope and asks for holy water, prepared to die. Franciscan and historian Donatus Mooney tells us that when O’Hea blesses the water, it transforms into milk, “sweet and healthy.” Upon drinking the milk, the invalid is miraculously healed.

 

1612: Despite repeated attempts by colonial administrators and settlers to destroy the fabric of Timoleague Friary and extract its valuable resources, local workers continuously repair the buildings. This is not without frustration. As soldiers destroy the friary’s alter decorations and glass windows, one carpenter threatens not to work anymore until and unless some divine retribution is enacted. The following day, Daniel O’Sullivan arrives with a small force and decimates the English soldiers, so that only one is left alive.

 

By the time of O’Hea’s miracle in 1595, Timoleague has long since been dissolved along with all the other monastic foundations in Ireland – at least on paper. In practice, it remains an active hub of social and religious activity. It survives through a constellation of supports: the continued patronage of lordly families in the surrounding region; the persistence of the friars who perform masses and sacraments despite very real threats of bodily harm and incarceration; and the repeated acts of care and repair by local workers and parishioners, who protect and rebuild the space as they are able.

 

We’re surrounded by catastrophe, both real and imagined, and generative AI is deeply implicated in it. Facial recognition powers state surveillance and police violence. Data centers rampantly consume and pollute natural resources, while companies flout regulations that should protect the marginalized communities in which they are built. Accelerating demand threatens to overwhelm essential infrastructures. Deep fakes undermine trust and facilitate online abuse. It’s an onslaught that potentially erodes our ability to cope and to resist.

 

Timoleague models a holistic approach to resilience and care in the face of catastrophe: care for community members experiencing harms to their physical and mental wellbeing; care for infrastructure damaged by greed and neglect; and care for the environment subjected to extractive practices. Like our anonymous carpenter, we are not obliged to bear continued injustice without frustration, anger, and resentment. But the community at Timoleague offers a reminder that through small, persistent acts of care, we might resist these growing injustices and the pervasive narrative that all of these things are inevitable.



Margaret K. Smith is Interim Director of the IRIS Center and Research Assistant Professor of Digital Humanities at Southern Illinois University Edwardsville, as well as a historian of medieval and early modern Ireland. Her historical work explores the relationship between law, landscape, and authority in Gaelic Ireland, and she has also published on topics in critical data studies, digital humanities infrastructure, and ethical collaboration practices.


The views and opinions expressed in this post are solely those of the original author/s and do not necessarily represent the views of the North American Conference on British Studies. The NACBS welcomes civil and productive discussion in the comments below. Our blog represents a collegial and conversational forum, and the tone for all comments should align with this environment. Insulting or mean comments will not be tolerated and NACBS reserves the right to delete these remarks and revoke the commenter’s site membership.

bottom of page