Aparigraha

Written By Kirthi Jayakumar

The AI economy has a word for taking as much as possible, as fast as possible, from as many sources as possible: Scale. Companies that have scaled the most have accumulated the most data, compute, mineral resources, intellectual property, and market share, are considered the most successful, most powerful, and the most important actors to center in governance. This logic of accumulation is deeply embedded in the culture and mindset surrounding the AI lifecycle, that questioning it is met with dismissal: Of course more is better, of course the goal is to grow! And yet, the idea of “growth” is defined on terms that benefit the already-powerful more aggressively. The benefits continue to accrue to a small number of people, as power continues to be corralled in their hands. The harms are vastly distributed, and disproportionately, at that. Aparigraha, the principle of non-possessiveness, offers wisdom to counter this reality.

Before we proceed, we want to state that Aparigraha is a living principle practiced across multiple South Asian religious and philosophical traditions, including Jainism, Hinduism, and Buddhism, each of which carries its own interpretation and emphasis. In Jainism, Aparigraha is one of the five great vows (Mahavratas) observed by both ascetics and householders. Here, it has its most systematically developed philosophical treatment. In the yoga tradition, it is the fifth among Patanjali’s Yamas, the ethical restraints that form the foundation of yoga practice. We draw on the philosophical dimensions of Aparigraha that have been articulated across these traditions while acknowledging their internal diversity. We welcome correction and deeper engagement from scholars and practitioners who carry these traditions.

Understanding Aparigraha

Aparigraha or the principle of non-possessiveness is found across Jain, Hindu, and Buddhist traditions across South Asia. The principle has been questioning extractive logics for two and a half thousand years, an approach that AI governance is still wanting: How much is enough and what does the actual taking cost for those from whom we take?

Aparigraha is a Sanskrit term, formed from 'a-' (meaning non) and 'parigraha' (grasping, seizing, possessing). It stands for non-possessiveness, or the disciplined refusal to take or accumulate beyond what is genuinely needed. In Jainism, aparigraha is the principle of limiting one’s possessions and desires, and applies equally to material and psychological objects, calling for one’s efforts to counter the attachment, expectation, craving, and extraction that accompany the drive to possess. The Jain text Samana Suttam makes the connection between possessiveness and harm explicit, by suggesting that attachment causes one to commit violence and theft, and to develop a will for unlimited hoarding.  

What Aparigraha unsettles in AI governance

Aparigraha poses a direct challenge to the data economy on which AI is built. The collection of personal data at scale, such as behavioral, health, locational, and communicative data among others, is the defining act of AI’s possessiveness. The industries that accumulate the most data about the most people are the most powerful actors in the AI ecosystem. Aparigraha calls on us to ask: What is the genuine need that this accumulation serves? How much data is enough for the legitimate purpose claimed? What are the harms to privacy, autonomy, and the communities whose lives are turned into training material that emerges from unlimited hoarding? While some may argue that data minimization calls on us to focus on the same things, it is important to understand that even data minimization is being observed as a framework of optimization, where the focus is on taking as much as is necessary for the purpose, rather than as a framework for restraint grounded in an understanding of what taking costs.

Relatedly, Aparigraha also poses a meaningful counter to the extraction of conflict minerals. Driven by an explosive surge in demand for hardware, semiconductors, and data center infrastructure, all of which rely on raw materials extracted from regions marked by armed conflict and human rights violations, this extraction of conflict minerals takes place with scant regard for the prevalence of conflict-related sexual violence, war crimes, crimes against humanity, and torture, all to keep up the mining operations going. Fairness in supply chains is neither prioritized before minerals are sourced, nor remedied after the impact of such sourcing comes to light. Aparigraha reminds us that we owe a duty of care by shifting from unlimited accumulation to mindful stewardship.

Aparigraha also challenges the intellectual property regime that governs AI. The accumulation of training data, model weights, and proprietary architecture under corporate ownership is a form of grasping that removes from the commons what was produced by collective human effort across generations. The knowledge encoded in language models was produced by billions of people across centuries. Corporations that own the models have grasped that knowledge without sharing the benefits of the grasp. Aparigraha insists that this arrangement is economically inefficient and morally incoherent, as the attachment to intellectual property as a form of competitive advantage is the kind of possessiveness that serves as the root cause of harm.

Finally, Aparigraha carries a specific practical wisdom for AI governance institutions themselves: The principle that those who govern should not accumulate power beyond what is genuinely needed for the governance purpose. AI regulatory bodies that seek to extend their jurisdiction, standards organizations that seek to capture the standard-setting process, and ethics boards that accumulate advisory functions across institutions all are subject to the same possessiveness that Aparigraha identifies as the source of distortion. A governance culture shaped by Aparigraha would regularly check to see what we are holding that we do not need, and what we would release if we were genuinely committed to the purpose rather than to the power.

References

This article draws from Jain, Hindu, and Buddhist philosophical traditions that have sustained the principle of Aparigraha across millennia. We are grateful to the scholars, practitioners, and communities who carry these traditions and have made their wisdom available for wider engagement. We invite readers to support organizations working on data justice and digital commons, which carry the Aparigraha spirit into contemporary governance contexts.

 

Kirthi Jayakumar
Next

Hózhó