@sujato
I deeply respect your perspective on the importance of preserving the human element in translating and engaging with Buddhist texts. The historical efforts of translators like you and others and others are indeed monumental achievements of dedication and faith, and I agree that no machine could ever fully replicate the transformative process of immersing oneself in the Dhamma.
However, I see AI not as a replacement for this deeply human work, but as a tool that can complement and enhance one’s interaction with the texts. History offers us similar examples: when photography was first introduced, many feared it would lead to the death of art. Instead, it lowered the barriers to creative expression, democratized access to new forms of art, and provided an invaluable tool for human progress and welfare. I believe AI has the potential to do the same for education as well as engagement with Buddhist teachings.
The concern about AI “exploiting” the work of translators and artists is understandable, but I see it differently. AI training is akin to a student learning from a teacher’s choices or an artist studying how a painting was constructed. It builds on the wisdom and efforts of others, not to diminish their contributions but to make their efforts and insights more accessible and impactful. AI is a tool, not a replacement for the lived experience and understanding that comes from practice and study.
Personally, I have found AI to be an invaluable educational aid. For instance, I’ve used AI to help answer questions about Pāli grammar, significantly lowering the barriers to engaging directly with the suttas. Rather than replacing study, it has enabled me to dive deeper into the texts and foster a more personal connection with them. Similarly, AI can serve as a private grammar or writing tutor, helping people overcome the practical challenges that often discourage engagement with Buddhist teachings. It is also has been an invaluable aid in formulation of thoughts and has in more than one occasion helped me practice more compassionate and understanding speech by correctly pointing out responses which may be interpreted as harsh.
I also believe there’s an opportunity for the Buddhist community to use and guide the development of AI tools specifically tailored to support practice. Open-source models are improving and specialized fine-tuned models could be designed to assist practitioners in identifying connections between suttas, navigating Pāli texts, or deepening their understanding of the teachings. These tools would not replace traditional methods of learning but would serve as aids to make the path more accessible to a wider audience.
The concerns you raise reflect an admirable caution and reverence for the traditions of Buddhist study, and maintaining a repository of human-translated materials is indeed crucial. However, I think there’s room for balance. Rejecting AI entirely risks missing out on its potential to unite and support practitioners. It’s not about surrendering to technology but about using it skillfully to foster greater access to the Dhamma. I personally feel that an outright rejection of AI on Sutta Central, while admirable in intent, may be counterproductive. Such an avoidance of AI is more of an inconvenience to those who sincerely wish to engage with the texts who do not have knowledge of the languages of the text on this site for which translations into their native language are available. I have had to copy many Chinese texts into ChatGPT and ask it to translate and give me the gist of the agama in order to engage with these materials. I understand the limitations of these models and I think a significant warning should be made regarding anyone who wishes to use AI translation in such a manner, but this practice has deepened my appreciation of the Agama literature.
If I may make a suggestion, it could be worth exploring the option of offering approved machine translations for texts that currently lack human translations—accompanied by a significant disclaimer about their limitations. These AI-generated translations could then be phased out as approved human translations become available. This approach would strike a balance between accessibility and the preservation of high standards for translated materials. The reality is that many, including myself, will continue using AI tools for such explorations, and this compromise would provide greater confidence in the accuracy and reliability of the outputs we are engaging with.
AI, like any tool, is only as beneficial as the intentions of those who use it. When integrated thoughtfully and ethically, it has the potential to support the transformative journey of learning and practice, not replace it. I hope this perspective offers a middle path between fear and trust in the face of this rapidly evolving technology. Thank you so much for your efforts and work Bhante Sujato!
