Washing machines and LLM AI: lessons from two grandmothers

Large language model technologies have the capacity to level the playing field in academe, but some will be opposed and want to preserve established hierarchies of knowledge and authority.

November 03, 2023
Une image de machine à laver et un panier de vêtements.

My two grandmothers held contrasting perspectives on modern ideals. Victoria, my German grandmother – who resided in the city – firmly embraced the modern narrative of progress, development and civilization, with white/European/German culture at its forefront, guiding us toward a future of prosperity, lawand order. Conversely, my Guarani grandmother, Vitalina, who lived in a humble countryside dwelling with a simple dirt floor and an open fire, viewed this same narrative as not only harmful to the land, but also dangerous for humanity and our shared destiny on a finite, living planet.

Their views on technology diverged as well, and surprisingly, Victoria, contrary to expectations, adamantly refused to use a washing machine. She believed that such machines diminished women’s value, as she saw women’s worth rooted in the quality of domestic service they could provide.

In stark contrast, Vitalina held a profound fascination with washing machines, but not for their utility in cleaning clothes. Instead, she was interested in the sounds, smells and motions. I have childhood memories of  watching together the top-loading washing machine complete its cycle, with the lid open (which was possible then). She encouraged me to mimic the motions of different stages of the washing cycle – like a dance, twisting my hips during agitation and rotating my hands and head around during the spin cycle.

Victoria’s opposition to the washing machine and her conviction that it devalued women’s inherent worth reflects a larger pattern of internalized oppression, where individuals may choose to uphold established power dynamics, even if it comes at their own expense. In academia today, a similar resistance can be observed in some objections to generative artificial intelligence, particularly LLM (large language models, like ChatGPT).

Broader concerns related to AI include the reproduction of biases, questions of equity, access and ethics, as well as the loss of human control and the appropriation of intellectual, artistic and cultural property. They also involve the depersonalization of learning, workforce vulnerability and other externalized environmental and social costs of AI, which arise from the resource-intensive nature of AI development.

These risks are substantial and should not be downplayed. However, in addressing them, we will encounter numerous paradoxes. One such paradox is that while AI can be employed to perpetuate existing colonial hierarchies and social inequities, it can and will also be harnessed to challenge and disrupt them.

LLMs are a case in point. For centuries, the written word has stood as a fundamental pillar of knowledge creation and scholarly pursuits, closely tied to the historical hierarchies of colonialism. Alphabetic literacy has functioned as a powerful instrument for ranking and classifying individuals, cultures and societies, giving rise to and enforcing notions of superiority and inferiority. The privileging of writing, particularly academic contexts, serves to perpetuate the colonial project of enforcing hierarchies of value and intellect. This approach aligns with the singular narrative of progress and human advancement championed by my grandmother Victoria and vehemently opposed by my grandmother Vitalina.

For many Indigenous students and others from historically and systemically marginalized groups, academic writing –often linked to colonial languages and modes of expression –has been a barrier to equitable participation in academia. Modes of assessment that privilege the written word and the English language simply cannot reflect the diversity of human intelligence, worldviews, modes of literacy and ways of knowing and expressing knowledge.

In this context, LLM AI technologies may offer a promising avenue for leveling the academic playing field. These technologies can recognize and accommodate diverse linguistic and cultural expressions, enabling students from oral, multilingual and multiliterate cultures and backgrounds, who have often been deficit theorized in academic contexts, to engage with and create academic content in ways that better resonate with their identities and worldviews. LLMs have the capacity to understand and generate text in multiple languages and modes of expression, thereby expanding access to academic visibility and recognition.

Certain criticisms directed at the incorporation of generative AI into academic research and education serve to perpetuate established colonial knowledge structures and hierarchies and may jeopardize access for marginalized groups to a level educational landscape. In a parallel to the washing machine’s role as a disruptor of traditional gender roles, LLM technologies have the potential to challenge entrenched norms and power dynamics within academia.

In essence, just as Victoria’s resistance to the washing machine reflected a desire to maintain patriarchal structures, the resistance to AI in academia may reflect an inclination to preserve established colonial hierarchies of knowledge and authority. Recognizing this parallel invites us to critically examine our choices and consider how generative AI can be harnessed to confront and challenge colonial legacies in knowledge production and dissemination.

In contrast to AI, washing machines do not present a threat to humanity. While I am cautious not to idealize Vitalina, her approach to technology perhaps offers the most valuable lesson for our interaction with generative AI. She engaged with the machine with an open mind, treating it as a living entity, driven by curiosity, and prioritizing the integrity of the encounter over the machine’s intended purpose. After a period of collaborative inquiry with me and the washing machine, Vitalina returned to her land, to her home, to her fire.

AI cannot replace connections to the land, home, or fire, even though it can skilfully simulate intimacy, sometimes deceivingly so. Nonetheless, if used with and for discernment and collective responsible stewardship, it does have the potential to be a generative relation, to democratize access to knowledge production, to enable new forms of research and collaboration, and to potentially redefine who holds authority in knowledge creation in academia.

In the spirit of full disclosure, as a speaker of English as a second language managing the demanding role of a dean in a faculty of education, I must acknowledge that I simply wouldn’t have had the time to write this article without AI assistance. While ChatGPT lacks biological grandparents, it helped me organize my ideas, suggested different analytical angles that I could have pursued and helped me refine the language of the text. We danced.

Vanessa Andreotti is the dean of the faculty of education at the University of Victoria. She is the author of Hospicing Modernity: Facing humanity’s wrongs and the implications for social activism and one of the founders of the Gesturing Towards Decolonial Futures collective. She identifies as a racialized settler on the ancestral lands of the Lək̓ʷəŋən (Songhees and Esquimalt) Peoples.

This site is registered on wpml.org as a development site. Switch to a production site key to remove this banner.