
A few years back, my dad was diagnosed with a tumour on his tongue – which led to some pivotal choices regarding his treatment. My family, rooted in a uniquely complex medical dynamic, faced a crossroads. My elder sister, a trained doctor, advocated for conventional medical interventions, whereas my parents leaned heavily on traditional remedies, reflective of our upbringing in a small town in India. This cultural background instilled within us a deep trust in holistic practices. My father’s ritual of obtaining herbal solutions from a local ‘vaithiyar’—a practitioner of Siddha medicine—was his way of maintaining a connection to familiar healing arts.
As the medical prognosis grew dire, the hospital doctors and my sister who were urging a surgical option met with staunch resistance from my parents, particularly due to concerns about potential impacts on my dad’s speech. It was at this moment I stepped in as the family mediator. Adopting the quintessential millennial approach, I leveraged the internet as a resource to bolster our case for surgery. Ultimately, after much deliberation and research, my father reluctantly agreed to the procedure. However, in a surprising twist, he evaded it by diverting our attention with family news, continuing instead with his herbal remedies. Remarkably, within months, the tumour shrank and ultimately vanished, stirring sentiments of triumph within our home.
At that time, I simply regarded the incident as an anomaly, a stroke of good fortune. But, in retrospect and amidst my evolving understanding of the digital landscape, I am led to question the dismissive stance I took towards traditional knowledge and the overwhelming authority of digital resources.
While pursuing studies on responsible AI system design at Cornell University, a disconnection became evident between my academic environment and the rich heritage I grew up with. This estrangement highlighted a significant issue: the current digital paradigm, exacerbated by generative AI (GenAI), reveals stark inequalities in knowledge dissemination. The foundational contents of the internet have predominantly reflected Western perspectives and, consequently, have marginalized non-Western philosophies and indigenous understanding—effectively creating a knowledge gap that could widen dangerously.
As GenAI becomes a primary learning tool for many worldwide, studies show that nearly half of user queries involve seeking practical guidance or knowledge. Even though AI systems claim neutrality, a closer examination reveals an inherent bias towards dominant knowledge forms, predominantly Western in origin, while sidelining alternative perspectives rooted in oral traditions and marginalized languages.
With each interaction, the perpetuation of these hierarchies through GenAI not only risks erasing complex systems of understanding established over centuries but also threatens the diversity of knowledge. The implications extend far beyond mere representation; they touch upon the resilience of societal wisdom itself.
The training of GenAI models on vast text datasets invariably excludes important oral knowledge. For instance, around 45% of training data is in English—spoken by merely 19% of the global populace—while major languages like Hindi contribute a fractional 0.2%. This situational disparity extends to Tamil, my mother tongue, which, despite a strong speaker base, occupies an even smaller data percentage.
The importance of language as a reservoir of experiences is paramount, shaping communities and their practices across time. Each language holds unique histories, spiritual beliefs, and ecological insights, often rooted in specific locales. When marginalized languages fade from the digital record, knowledge intricately tied to them risks disappearing.
For example, local ecological comprehension often hinges on the specific nomenclature of regional flora. An environmental acquaintance highlighted how a community’s connection to nature is identifiable through the terminology they possess for local plant species, underscoring a loss of knowledge with language erosion.
This sentiment was echoed in discussions with experts such as Dharan Ashok, whose work on reviving traditional building techniques in India emphasizes the intersection of language, ecology, and Indigenous knowledge. With knowledge often orally transmitted through local dialects, the death of an elder can herald the loss of invaluable information.
Antonio Gramsci’s concept of cultural hegemony provides a framework for understanding how specific epistemologies dominate, achieving an unrecognized global standard that obscures the historical circumstances of its ascendance. Western paradigms are commonly viewed as neutral, universal truths, thanks to institutional endorsement through academia and formal organizations.
The environmental implications manifest starkly in the architectural designs that emerge from these dominant frameworks. For instance, modern glass-clad buildings, while facilitating efficiency in cooler climates, exacerbate overheating issues in warmer regions, transforming them into burdens on the environment. Within places like Bengaluru, the city is reeling from a water management crisis that further illustrates how traditional, community-based ecological practices have been replaced by centralized, often unsustainable solutions.
Experts are increasingly highlighting that as AI technologies charge forward, there exists an alarming trend of homogenization in knowledge access. With AI-generated content flooding the internet, there is a risk of generating a feedback loop wherein dominant ideas overshadow niche or alternative knowledge streams.
Andrew Peterson’s notion of “knowledge collapse” captures this phenomenon; a narrowing of human accessible information that neglects lesser-known perspectives while increasingly relying on the convenience of popularized content. As educational systems entwine with AI, future generations may find themselves distanced from rich cultural heritages tied to local practices.
While some may argue that enhancing datasets could resolve the knowledge disparity, the challenges of inclusive representation and understanding extend well beyond data collection. A senior leader working on an AI initiative for millions of farmers in Asia reflected how many pristine local practices remained excluded due to a heavy reliance on mainstream scientific literature, revealing the underlying tensions between risk mitigation and comprehensive knowledge practice.
Organizations dedicated to preserving Indigenous knowledge face a turbulent path, often met with systemic barriers and skepticism regarding legitimacy, further exemplifying the institutionalized marginalization of local practices recognized for their value in their respective contexts. I was privileged to work with a nonprofit aiming to legitimize Indigenous knowledge in agriculture, navigating the treacherous waters of validation amidst a sea of scepticism.
The lessons learned through these experiences reveal that the erasure of local knowledge is not merely a tragedy for those communities; it symbolizes a global crisis with heart-wrenching ramifications for ecological and societal well-being. We are woven into shared ecosystems where the health of local knowledge systems directly impacts global sustainability.
Writing this reflects a deeply personal journey wherein I grapple with my own uncertainties about traditional knowledge, even as I advocate for its inclusion in contemporary discourses. The path forward demands that we cultivate curiosity beyond superficial acknowledgment and critically engage with diverse knowledge systems rather than dismissing them. As we advance into an AI-dominated future, the questions remain: Are we prepared to embrace the wisdom embedded in long-standing knowledge systems while safeguarding our ecosystems, or will we perpetuate a cycle leading irrevocably towards a homogeneous and ultimately impoverished understanding of the world?
It seems acknowledging my uncertainties—whether concerning my father’s herbal tinctures or deeper knowledge hierarchies—might be the honest starting point as we navigate the complex terrain of knowledge in this AI-era.
.