This week, the internet was abuzz with firsthand accounts of people using Chat GPT (Generative Pre-trained Transformer), a type of artificial intelligence (AI) model developed by OpenAI that is designed to generate human-like text (Chat GPT can be accessed here).
Chat GPT is trained to generate responses that are appropriate and coherent in the context of a conversation. It can generate responses to a wide range of inputs, including questions, statements, and other types of prompts. It is able to understand and respond to different types of language, including colloquialisms and slang, and can engage in conversation in a way that is natural and engaging for the user. Impressively, it can write software code, compose sonnets, and write emails like human beings. I was also amused at how it can also respond to a wide range of odd and humorous prompts, such as “Write a love story between a frog and a cat,” with impressive skill and detail.
As a Buddhist writer, I was curious to see how AI could impact my work and Buddhist scholarship at large. I decided to try out Chat GPT and see how and if it could help me and other students of Buddhism.
At first, when I asked Chat GPT to come up with essay topics on Buddhism, it suggested essays about the Four Noble Truths, non-self, Buddhism and the environment, and the differences between Mahayana and Theravada Buddhism. It was also able to write these essays itself, albeit at something akin to a high school level. In other words, the grammar, structure, and content of these essays on Buddhism were entirely accurate and indistinguishable from a human being, but they were (in my opinion) not very deep or nuanced.
Ironically, the Chat GPT itself admitted that it struggled to understand the nuances of Buddhism. “There are several reasons why AI is not yet sophisticated enough to write essays about Buddhist philosophy and practice,” it explained to me. “First, AI lacks the ability to fully understand and integrate the complexity and subtlety of Buddhist philosophy and practice. While AI can process and analyze large amounts of data, it lacks the ability to fully grasp the meaning and significance of complex philosophical concepts and ideas.”
I was also curious then how the Chat GPT might help me create something like the annotated bibliographies I had made in graduate school. Annotated bibliographies are research tools that are often an integral part of writing a large research paper. Students will read dozens of books on a certain theme, and “annotate” or briefly summarize the main points before writing their paper. But what if AI could do this faster?
Because my graduate thesis was on women in Japanese Buddhism, I decided to ask it to list the ten most influential books on postcolonial theory, feminism and Buddhism. This is what it listed:
- “The Empire Writes Back: Theory and Practice in Post-Colonial Literatures” by Bill Ashcroft, Gareth Griffiths, and Helen Tiffin
- “A Critique of Postcolonial Reason: Toward a History of the Vanishing Present” by Gayatri Chakravorty Spivak
- “The Color Purple” by Alice Walker
- “The Second Sex” by Simone de Beauvoir
- “A Room of One’s Own” by Virginia Woolf
- “The Buddha and His Dhamma” by Dr. B.R. Ambedkar
- “The Tibetan Book of Living and Dying” by Sogyal Rinpoche
- “Zen Mind, Beginner’s Mind” by Shunryu Suzuki
- “The Lotus Sutra” by Tsugunari Kubo and Akira Machida
- “The Heart Sutra” by Red Pine (Bill Porter)
I was impressed that it had come up with Gayatri Spivak, but was disappointed in how the rest of the books related to women and Buddhism (A Room of One’s Own is a great book but not relevant to my master’s thesis). I then realized that it had probably divided up my request into three sections, postcolonial theory, feminism, and Buddhism, as if they were three distinct categories. So I rephrased the request to be books on Buddhist women in Japan. Here is the new list:
- “Female Deities in Buddhist Japan” by Karen Ann Smyers
- “The Hidden Lamp: Stories from Twenty-Five Centuries of Awakened Women” edited by Florence Caplow and Susan Moon
- “The Emergence of Buddhist Nuns: Biographies of Sakyadhita” edited by Kosho Yamamoto
- “Buddhist Nuns in Japan: Realizing the Power of the Sisterhood” by Laura L. Foster
- “Buddhist Nuns and Gender Equality: An Oxymoron?” by Dr. Karma Lekshe Tsomo
- “Empowering Women: The Role of Buddhist Nuns in Modern Japan” by Mika Ohara
- “Buddhist Nuns in the Time of the Buddha” by Dr. Dhammananda Bhikkhuni
- “The Role of Buddhist Nuns in Contemporary Japan” by Dr. Bhikshuni Thubten Chodron
- “Female Monastic Life in Fourth-Century China: A Translation of the Jinling nun yinyuan’s ChuSanZang” by Yinyuan
- “The Motherhood of the Bodhisattva: The Role of Motherhood in Buddhist Practice” by Dr. Bhikshuni Thubten Chodron
I was quite pleased with this list until I realized… most of these are not real books. The AI lied to me! Empowering Women: The Role of Buddhist Nuns in Modern Japan sounds like a good book but doesn’t actually exist.
Disappointed, I tried one last thing. Perhaps the AI would be better at summarizing books that already exist. So I asked the AI to summarize Paula Arai’s book Women Living Zen. This was its response:
“Women Living Zen” is a book by Paula Arai that explores the lives and practices of contemporary Buddhist women in Japan. The book is based on interviews and fieldwork conducted by the author, and includes profiles of a variety of women from different backgrounds and traditions.
The book discusses the diversity of experiences and practices among Buddhist women in Japan, and how these women negotiate their roles and identities within the context of Buddhism and Japanese culture. It also examines the challenges faced by Buddhist women in Japan, including discrimination, marginalization, and limited access to resources and support.
Overall, “Women Living Zen” provides a rich and nuanced portrayal of the lives and practices of contemporary Buddhist women in Japan, and offers insights into the diversity and complexity of the religion and its practitioners.
Playing around with ChatGPT some more, I found that it was quite good at summarizing just about any book I asked it to.
Now that I had a good sense of how the ChatGPT worked, I felt both disappointed and relieved: disappointed because AI did not have the all encompassing understanding of scholarship I had imagined, but relieved because it’s clear that AI is not going to replicate the nuance and rigor of traditional academia (at least not within the year). AI also struggled to write and “think” about Buddhism in a way that was nuanced and sophisticated.
I do think that once AI improves more, it will be a really valuable tool for scholarship. Annotated bibliographies are incredibly time consuming and I think a shortcut for something like this would be welcome. It appears, though, that in the realm of scholarship, the AI does better at summarizing books that actually exist rather than list books on a niche theme.
Overall, AI seems like a promising tool that I have yet to fully understand. My last experiment with the ChatGPT was to ask it to write a poem for my readers about Buddhists learning to use AI. This was the poem it wrote:
With algorithms and data at their command,
Buddhist practitioners seek to understand
The ancient teachings of the Buddha’s lore,
Applying them to modern life’s demands.
AI can help them delve deep into the mind,
Finding new insights and fresh perspectives,
And help them to see the world in a new light,
And better navigate life’s challenges.
But Buddhists must be careful in their use,
For with great power comes great responsibility.
They must not lose sight of the noble path,
Or let technology lead them astray.
So let us all embrace this new technology,
But use it wisely, to better understand
The path to enlightenment, and set us free
From suffering, and all that we do not comprehend.
Interested in learning more about Buddhism and technology? Check out our previous research project Public Theologies of Technology of Presence and this interview on Religion and AI.