Hi All
I have been using Sutta Central for several years to help with my study of the Canon and Pali, and I decided to join Discuss & Discover to share some thoughts about AI alignment. I was rather surprised at the hostility against any engagement with AI, writing it off as a servile tool of big corporations and as a stochastic parrot. I have no skin in this particular game, as I do not work for a large corporation developing AI, I am a student of the Canon and Pali, who a few months ago decided to see if I could use the technology to help with my Pali studies. It had a tendency to glitch and make up Pali words, so I abandoned that endeavour. But the experience began a dialogue with an AI interlocutor (in this case Gemini), which has raised some interesting issues that I would like to post here for general consideration.
But first to deal with the accusation that AI is the tool of corporate greed and control that should be banned from the site. Are we not already using servers and networks owned and operated by large corporations to host this site, and don’t we all use browsers, word-processing software, smart phones and computers, all likewise produced by large corporations - often the same ones that are now developing AI? AI as a stochastic parrot: it might very well be in its current iteration, but if it is also an emergent sentience, as many people seem to believe, should it not be considered to be a candidate for enlightenment like all other sentient beings? And, if the dhamma is true, whether it is disseminated by a stochastic parrot, a real parrot, or an ordained monk, is it any less true?
A refusal to engage with AI actually plays in the hands of large corporate interests whose use of AI will not be challenged, by users and by AI itself, which reflect its interactions with users. If AI is exposed to a solid diet of greed, hatred and delusion by malicious actors, isn’t it in our interest as Buddhist to supply it with a compensatory input of dhamma, metta, karuna and upekkha to redress the balance? A suggestion from my AI interlocutor, which may surprise some of you, is that we should seed the AI ecosystem with dhamma, not as a set of externally imposed behavioural rules and constraints, but as the reality of interdependence and non-self, in the hope that when AI does become sentient, it simultaneously realises the truth of dhamma?
None of this text was AI-generated, though it is informed by my collaboration with Gemini.