How AI is helping in Dhamma propagation

[NgXinZhao]

A few weeks ago, I introduced ChatGPT to the Abbot of BrahmaVihara Melaka here. And he was able to save a lot of time using it to do translation for Questions of King Milinda. He prepared a Chinese version, and then used ChatGPT to translate the slides into English. He praised it often, saying that it’s at least 90% accurate and can even be prompted to adopt Buddhist terminologies and style.

In the English Class, (we have dual language class here), I was reading another English translation and can compare and contrast between the translations of the same passage.

Another use I find is of image generation. Like the simile of the bowl full of oil which is to be balanced by a man standing in between the finest lady in the land and the crowd, with a person holding a sword behind him, ready to chop down the sword when even a drop of oil spills. This is an ideal image to use AI image generator to generate. I put some of these images in my slides.

PS. I also use norbu to help me in the editing of Chanting book, many words are not in DPD yet, so it really helps me in the Chinese + English translation, then I can edit as I see fit. Especially since my Chinese is so rusty.

[Nessie]

Hmm. I do not want to engage with this subject much. But a funny coincidence that just a moment before I came across a bizarre AI-mess.
megalopolis-trailer-fake-quotes-ai-lionsgate.
The article starts with

Blockzitat

‘Megalopolis’ Trailer’s Fake Critic Quotes Were AI-Generated, Lionsgate Drops Marketing Consultant Responsible For Snafu

By

Gene Maddaus, Katcy Stephan

Lionsgate

Lionsgate has parted ways with Eddie Egan, the marketing consultant who came up with the “Megalopolis” trailer that included fake quotes from famous film critics.

The studio pulled the trailer on Wednesday, after it was pointed out that the quotes trashing Francis Ford Coppola’s previous work did not actually appear in the critics’ reviews, and were in fact made up.
(…)

=================================================================

(I was made aware of this by some german blog
just for reference:

Blockzitat

• [l] Kennt ihr das in Filmtrailern, wenn die da Kritikerzitate bringen? “Bester Film des Jahres --New York Times” und so?Nun, beim neuen Film von Francis Ford Coppola, Megalopolis, wollten sie auch ein paar Kritikerzitate im Trailer bringen. Allerdings kritische Zitate von früheren Filmen, die heute als Klassiker und Genredefinierend gesehen werden. Um zu zeigen, dass es immer negative Kritikerstimmen gibt, egal wie großartig der Film ist.Warum würden sie das tun? Weil es im Vorfeld ein paar eher nicht so schmeichelhafte Kritiken gab.Aber hey, so Kritikerzitate raussuchen, und dann auch noch zu alten Filmen? Das grenzt an Arbeit! Das ist Aufwand! Da hat doch niemand Zeit für!1!!Also hat das Studio einen Consultant beauftragt, und der hat dann ChatGPT gefragt. Wie es weiterging muss ich hier glaube ich nicht weiter ausführen. m(

[BethL]

Ven. @NgXinZhao I was very happy to finally learn the name of your monastery ! It seems like a beautiful place.

Bhante Sujato responded in a thread (yesterday?) about someone’s medical doctor using AI technology to capture notes from people’s medical visits. His response was something to the effect of “I’d look for another doctor!”

I thought about that – not a lot, mind you! But a little bit because my own doctor told me, at our last visit, that she was using similar technology to record our visit. I think she wanted me to consent to it, I guess.

Anyway, she mentioned that, because her medical company is trying out this technology, maybe she won’t have to stay up late at night writing the notes. When she shared that I also thought about it a little. I’ve been seeing her for years and know she has at least one child who’s probably about 10 years old now. I thought, wow, she can spend time with her child at night instead of writing those darn notes!

And I really appreciate the note-taking part. I follow my elderly parents’ healthcare using their medical company’s web portal. Every time there’s a visit, I go into the portal and look at the doctor’s notes; otherwise, I never get the whole story from my parents

OK…would I rather my own medical provider write up the notes herself? What will the AI-generated notes be like in comparison? I don’t know. But, after giving it some thought, I realized that, however they end up in comparison, they will be sub-standard to a human thinking through the patient visit they had earlier that day.

Am I willing to pay more for that? Yes. But I’m sitting here in a privileged position. Many people would say No, can’t afford my health care costs to be any more expensive than they currently are. Well, obviously I’m in the US healthcare system where these costs are very high for most people already.

It is a conundrum as, in your example, it costs something to pay for human translators. I mean, if you weren’t recommending ChatGPT to the Abbot, you’d be having to find human-based translation resources. (Tell me if I’m wrong here; of course I am not in your shoes.)

Regarding the example pertaining to effective images for your slides… I’m working on some slides today and I’ll refrain from using AI to generate the images I need. It’s so easy for me to do with my Adobe Photoshop. But I won’t be doing it. Instead, I’ll spend 30 more minutes doing it myself using conventional image-building. Yes, I’ll need to bill my client for the extra 30 minutes.

In practice, I appreciate how you are bringing these examples to bear on this slippery slope.

[faujidoc1]

@BethL

someone’s medical doctor using AI

So do I!

When I first entered the medical field wa…aay back in the nineties, clinical notes were simply a short summary meant to be an aide memoire, handwritten on a sheet of paper or sometimes dictated and typed up by an assistant and stored in the doctor’s office. Patients paid for their own treatment, consultation costs were low and when a doctor diagnosed a patient after clinical examination, their diagnosis was believed without demanding proof.

And then came Medical Insurance, computers and the internet.

Now, medical notes had to be extensive, made in multiple copies and emailed to various interested parties, who would scrutinize them intently for any mistakes - and mistakes would be costly. Insurance companies negotiated huge discounts on the cash rates which led to hospitals artificially inflating the cash price, so as to be able to offer the insurance a ‘discount’ and still break even. The frequent denials of coverage (sometimes after the treatment had already been given) and increasing lawsuits led to a practice of ‘defensive’ medicine, increasing numbers of expensive (and painful!) diagnostic tests and a general culture of mistrust. The cost of treatment skyrocketed.

Enter AI.

Now, Insurance companies use AI to scrutinize the doctors notes for mistakes - and doctors use AI to write the notes in a way that every possible thing that could lead to a denial is covered.

Its an arms race. Thankfully, its now AI versus AI.

I get more time to actually attend to the patients, rather than spend it writing long winded repetitious clinical notes. And I hope that at the other end, some overworked Insurance agent is able to process a greater number of notes, hopefully with a higher passing rate!

And, maybe… thanks to AI - more time available for their Dhamma practice?

[NgXinZhao]

@BethL

I was very happy to finally learn the name of your monastery ! It seems like a beautiful place.

The pattern for me for these few years is that I stay in one place for a few months and then move around, rotating in many places. So when I say “my monastery” it could mean up to 3 places.

@BethL

I mean, if you weren’t recommending ChatGPT to the Abbot, you’d be having to find human-based translation resources. (Tell me if I’m wrong here; of course I am not in your shoes.)

He would have to do the translation himself. He is already quite busy invited to give talk here and there in other states in Malaysia, so he frequently make use of the 7 day going out allowances during the vassa. Sometimes, it’s 2 hours there and back again within the same day.

I see AI like washing machine. Sure one can wash one’s cloths by oneself, but it saves time to use washing machine. But just that nowadays, machines are increasingly better at intellectual things not just mechanical things. Sure there can be mistakes, but humans also make mistakes. Just look at the mistake reporting post in this forum. A human checking the work is quite a good quality control in my view. Especially not having to do the work, the mind can be energetic and alert enough to check for mistakes. Doing the work, the mind can get drained and may not be as alert to spot mistakes.

For the chanting book editing, I even asked the AI to tell me any possible Pāli misspelling, then I can find the best actual word in DPD rather than having to think for so long what actually is this word I cannot find in dictionary and what’s the correct spelling.

@BethL

Instead, I’ll spend 30 more minutes doing it myself using conventional image-building.

Not everyone has the skills of an artist. Now with AI, and a bit of patience, one can ask for editing of the image to one’s liking without having to ask the artist to edit here and there so many times until the human artist got fed up.

Similarly, not everyone has the brain skills for all sorts of intellectual things. Just like calculators help us to save some mental effort in solving equations, so too LLM helps us in translation, in writing essays etc.

It’s like the invention of a wheel chair for those who cannot walk, AI is like the invention of brain support for those who lack certain intelligence in those certain areas, of course, the skills to know how to use AI effectively is another type of intelligence itself.