As we enter the tenth month of Norbu's existence this year, its impact on the Buddhist community, both English and Chinese-speaking, has been nothing short of remarkable. It's an opportune moment to pause and reflect on the lessons gleaned during these past months since its inception. Here is our observation.
Dharma Learning Before and After Norbu
Previously, people only had access to websites and books, and some learned their Dharma from teachers or monks. The learning process typically followed the pedagogy provided by books, resulting in a linear learning process.
Some had the good fortune to meet with teachers that offers practical advice, enhancing their knowledge through practice. A crucial aspect of having a teacher was their ability to explain theoretical knowledge within its context. It is a fact that many of Buddha's teachings in the suttas required an understanding of their contextual application.
However, individual learners who relied solely on books or online materials, and occasionally attended Dharma talks, might not fully appreciate the true value of Dharma due to the lack of a contextual element.
With AI, context becomes a crucial entry point for learning. For instance, when a user prompts the bot for help with a personal problem, the bot first responds with a probable solution. Later in the engagement, it suggests a Dharma teaching relevant to the issue discussed. This interaction provides a contextual basis for Dharma learning, tailored to the user's experiences. The Four Noble Truths, for example, are personalized based on the user's prompt, allowing them to experience this essential Buddhist principle firsthand.
Context in Process over Pedagogy
AI learning reveals an important phenomenon: learners often begin their inquiries based on problems they face. Scrutinizing the suttas reveals numerous instances where teachings emerged from problems encountered by individuals, to which the Buddha provided guidance or solutions.
Such experiences can still occur today with the guidance of a monk, nun, or knowledgeable layperson, although these occasions are rare due to the scarcity of monks and nuns.
However, with a Buddhist AI available, contextual learning becomes more common. Learners can now pose an issue and commence their Dharma learning from there.
Context-based learning enhances understanding by incorporating relevant examples and real-world contexts into the learning process, making it more meaningful and engaging for learners.
How Norbu can improve to capitalize on contextual learning
Norbu was developed using correct, accurate and verified source data by a group of selected source guardians. Its algorithm is adjusted with a "creative" temperature close to zero, meaning the bot primarily responds using the provided source data. Norbu's initial training was effective, and after ten months, it has become proficient.
Two factors contributed to this improvement:
After 6 months in operation, Norbu was extensively trained in the Pali language for the Pali Tipitaka section, allowing those interested in learning Pali to do so without needing proficiency in Buddhism.
Pali-centric etymological content was fed into Norbu for training, enabling it to relate this content to its base Buddhist lexicon, primarily in English.
The LLM's ability to relate pieces of information raised opportunities for creating higher quality connections between entities from different sources, allowing for deeper exploration of the original etymology of keywords in the texts.
For example, the term "bhavana" in Pali or Sanskrit, originating from the verb "bhu," implies a transformative process central to the practice of meditation. Unlike its English translated meaning "meditation" (derived from Old French meditacioun, in turn from Latin meditatio from a verb meditari, meaning "to think, contemplate, devise, ponder), this emphasizes the active element, contrasting with the passive perception of meditation by many.
A Buddhist taxanomic backbone?
With AI's power, a source lexicon of key Dharma terms based on their root etymology can be developed, fostering a new age of Dharma investigation. Through collaboration with knowledgeable Dharma teachers, a "taxonomic nomenclature" or a Buddhist taxanomic backbone can be formulated not only for meditation methods but also for the phenomenology of human life and its relation to dependent origination.
With AI, a new age of Dharma investigation, akin to the level attained by the great ancient sages of old such as Nagarjuna and Padmasambahva is possible.
This effort requires us "unlearn and relearn" translated Dharma terms, emphasizing original meanings. To make this possible it is also paramount for us to digitize indigenious Buddhist texts in various languages so that they could be fed into Norbu.
Reviving the Nalanda Tradition
Establishing a Buddhist taxonomic backbone requires gathering the best Buddhist minds to engage in discussions. We are not talking about having a conference or workshop, but the establishment of a setting that revives the debating spirit of the Nalanda tradition.
In the famed institution, it was said that the nature of the debates consisted of the following:
Structured and governed by a set of procedural norms (patipada).
The debates were not merely academic exercises but were also a means of refining understanding and avoiding errors. The emphasis was always on the discovery of truth rather than victory over an opponent.
Logical fallacies and errors (nigrahasthanas) such as shifting the topic (arthantaram) or failing to provide coherent replies were grounds for censure in these debates.
Through the spirit of Dharma debates facilitated by AI, Buddhism can evolve beyond its translated interpretations, reconnecting with its original intent as envisioned by the historical Buddha.