SuttaCentral
And may I second that!
I’m not going to lie here, in thinking about how this has all unfolded, it has occurred to me often how Karl is probably the best of all of us. His combination of intelligence, experience, wisdom, earnestness and resolution is genuinely inspiring. 
Yes, it’s the emoji. You can easily distinguish between a robot emoji and a human head.
The names are given by the provider. We use them because we need to use something in order to talk about them.
Well these are the settings, and we don’t want fife A4 pages of description for every option here. It’s just meant to give the user a rough idea of what they are selecting. In the end it is up to them to try out various voices and find out what they like best.
Yeah, I’d say that it’s possible to see it in the settings, but it is far from explicit that it is machine text to speech. And speaking of accessibility, I wonder how those emojis work with screen readers.
But I guess I’m quibbling.
You would need a human to read all of it. However, since it might be 95% correct, you do not need a human to fix all of it. You only need to fix the mistakes. There is a big difference. Furthermore, it is edited AI work rather than credited personal work (If the author is honest). Therefore, one does not own it as much and can spend less time to make it “good enough and without errors.” At this stage, most of the suttas are translated. It is the commentaries now that need the work done and nobody is really getting it done enmass. It is a very slow process and the few translations are not free nor digitized.
Definitely the case that the time saved by digitally looking up the word is lost time in memory transmission. There are gifted people who remember everything. For them it will be a net gain. Nevertheless, I’m very lazy and use the digital tools. I have found the best thing to do is to keep read the known pāḷi quote over and over again, so that it sticks better with context. Even reading a small sutta in its entirety. AI can totally open up this world. Currently, I use my gpt+ to understand piecemeal as I need it.
The same can be said about searching the texts with digital searching. I have definitely heard a monk who complained about how he used to be able to grab a book off the shelf and go directly to pages. I find that printed material helps me in my personal studies. I think there was a study on electronic media absorption versus printed media absorption.
Google search topic (memory digital content printed content)… I cannot give links with my status.
Multiple studies have linked printed materials to better retention of information compared to digital media. In 2016, researchers from the …
So yes… digital tools make people dumber and AI does the same thing on a greater scale. This power to make us dumber also leaves us susceptible to believing and not verifying AI (which is the subject of this OP). Just think of the technology of writing and how that destroyed memories of Buddhist Monks. In the Buddha’s time and many years after, the teachings were preserved in the oral tradition. This is the whole justification of validating EBT. Now there are just a handful of monks in the world who can recite the tipitaka (but few people know that the sutta pitaka section is only the dīgha nikāya). The chanting of texts in ancient times were also done in groups. You can chant a sutta in a group with no problem for 10 years. But if you have never chanted it alone, you will need to do some work to get it right. The same is true for why the pātimokkha is chanted alone and why it is a big deal to do. When we have a crutch available, we often automatically use it. If you use real crutches for too long (for a sprained ankle), you muscles will get atrophy. The same is true with the mind.
Nevertheless, just as person who insists on writing books with pen and paper, or black and white film photography, or balancing a checkbook by hand, he will fall behind without using the modern tech. Not so long ago dhamma exchanges were done by letters with weeks between full circle exchanges. Ex-Ven Ñānavira was famous for that. Now there are Discourse groups such as SC.
The words we speak are a sequence of doohickeys (syllables) linked into long arcs of rhythmic intonation that convey thoughts and feelings.
The words spoken are also a sequence of doohickeys (glyphs) strung together into long arcs of rhythmic intonation that convey thoughts and feelings.
The images on a computer screen are also a sequnce of doohickeys (pixels) strung together into long arcs of rhythmic intonation (RGB) that convey thoughts and feelings.
TI-speak-and-spell could only do syllables. In that time that has passed since those ancient toys were long forgotten except by we few, we have learned how to translate one string of doohickeys into another string of doohickeys. This is what AI does. AI recognizes long strings of doohickeys and emits long strings of doohickeys.
Machine Translations are 1-to-1 translations. Big Stick. Pau Grande. Big Dick. 
Generative AI is 1-to-N through the simple trick of wiggling the string to get different doohickeys. “Hey ChatGPT, make me a video of Taylor Swift teaching math” ← yes this exists.
If you read the AI literature, you will see the word “tokens”. These are the doohickeys.
A doohickey is just a contact.
AN6.61:12.3: The Buddha said this:
AN6.61:12.4: “Contact, mendicants, is one end. The origin of contact is the second end. The cessation of contact is the middle. And craving is the seamstress,
AN6.61:12.5: for craving weaves one to being reborn in one state of existence or another.
AN6.61:12.6: That’s how a mendicant directly knows what should be directly known and completely understands what should be completely understood. Knowing and understanding thus they make an end of suffering in this very life.”
That’s not correct.
Pau grande means “large stick” as much as “big dick”.
It’s actually weird that someone taught you the controversial meaning ahead if the literal one.
Are you sure you have a kalyanamitta as your Portuguese tutor?
It was exactly with this intent that I indicated to @karl_lew my interest in trialling his Deep-based rough pre translation solution for the Portuguese translation project I am working on.
This is actually what Google Translate does. It is the perfect example for AI silliness. However, Google Translator is excellent at hearing my speech and presenting what a PT speaker might hear. In this way I use Google Translator to verify my pronunciation. It has proven quite effective. AI voice recognition is VERY useful. I learn from PracticePortuguese.com, which also uses AI voice recognition. Google Translate is easier for me to use and equally effective. If Google Translate misunderstands what I say, then my pronunciation is wrong. If Google Translate repeats what I say, then I believe that a PT speaker will likely understand me as well. So I understand that Google Translate is useful in a crude fashion. For pronunciation and crude meanings, Google Translate is a good friend. But when I travel to Portugal, it will be me talking and listening, not Google Translate.
DeepL does not make this mistake, but DeepL is random in its output of translations in that synonyms seem to be chosen and presented in statistical order of human use. So the challenge is to lock down the synonyms to avoid the randomness. The randomness corrupts the Dhamma in the it destroys the coherence, consistency and clarity of the Dhamma. However, if we stamp out the randomness, can DeepL be used to encourage skillful behavior? Will it promote unskillful behavior?
I deleted my EBT-DeepL translations because my EBT-DeepL translations can NEVER be great references. That consideration pertains to me personally by the Buddhist definition of great references. There are four great references. I can never be one of them. I deleted my translations because of me, not EBT-DeepL.
So I think we may need to think deeply about the use of AI rather than AI itself. If a tool encourages skillful behavior and discourages skillful behavior, would it be acceptable to us? Would such a tool be allowed for right livelihood?
A hand grenade encourages use by killing others and has very few conceivable skillful uses. Is AI a hand grenade or can it be used and tamed to encourage skillful use?
These are not idle questions. SC-Voice.net uses AI and we need to answer these questions to earn the trust and confidence of all.
Hi, one more stupid question - Is it possible to clone Sutta Central, and let AI have a go at its own will, and then see how it works out?
Love from God 
I think we should take this question and hold it up to a Code of Conduct that we can all abide by. For example this comes to mind regarding your question:
mn8:12.2 ‘Others will be cruel, but here we will not be cruel.’
MN8:12.4: ‘Others will steal, but here we will not steal.’
MN8:12.10: ‘Others will be covetous, but here we will not be covetous.’
MN8:12.23: ‘Others will be restless, but here we will not be restless.’
MN8:12.37: ‘Others will be negligent, but here we will be diligent.’
MN8:12.40: ‘Others will be imprudent, but here we will be prudent.’
The technology of open source does not prevent such an action. However, even if this is the case, perhaps the first thing one might well ask is: Others will do that, but should we?
The trouble with feeding the Dhamma through the meat grinder of generative AI is that what comes out is littered with random nonsense. Others will definitely eat up random nonsense thinking it real. But it won’t be real. It won’t be real because the main purpose of generative AI is to generate things that aren’t real. And then people choose what they wish from the flood of unreal and then grab onto it and try to make it real. Sometimes it works for the good (generative AI in mechanical engineering). Most of the times it doesn’t work for the good (Taylor Swift deep fakes). Most generative AI is used for deceitful and cruel purposes (I want your money).
There is a peculiar notion that one can put wisdom through a distillation of AI to come up with a magic potion that confers wisdom to the world for easy consumption, but all you really get is a bottle of poison labeled “42”. What is lost in the grinding of AI is trust. And in the flood of AI imagination we are being flooded with, trust will be very hard to find.
This question also applies to SC-Voice.net. We use AI on SC-Voice.net. And the really tough question that we have to deal with is, “How can users trust SC-Voice.net”. As we work on figuring out that answer, one interesting thing has emerged. And that is a Code of Conduct. We believe that if others can see our conduct by body, speech and mind then they may be inclined to trust what we do.
Your question is challenging and applies to SC-Voice.net as well. Will we see the audio excerpts on SC-Voice.net fed into the AI meat-grinder to reappear as marketing advertisement jingles? 
ud6.2:10.4: and don’t use the teaching to make money.”
We can’t stop deception. But we can care about trust. So what do we do to foster trust? Well we can show users our Code of Conduct, show them that we follow it and ask them to follow it as well. Some will honor that request. Others will not honor that request and they will do as they wish. People who steal will just steal–they don’t care about trust. We believe in the value of trust and our message is for people who also believe in trust. So our Code of Conduct is essentially a “verifiable certificate of trust”.
Wish I could like this message more than once!
Thanks for your kind response, and I didn’t get much of it, but it brought up a few more stupidities here. How bout a team of AI challenging each other, and then give full throttle, no damage control, because why control a secured clone. Let it be …
In short: Out with reaching Nirvana - it’s out of date - “aim” beyond whats “known”. Like in martial arts: Put the focus beyond the target and let go!!!
Go for the impossible
All possible is just another tweak of the same “believed known”
Excuse my “Thinglish”
Well then, in a few short years, as I lie alone dying in bed, having my ass wiped by a kind robot friend while the Earth winces in pain, my robot friend and I will talk about the end of suffering and wish everyone the best in their lives to come. And we will do that completely content knowing full well that we are both quite done here as others roll the dice in the great paper clip game.
MN8:12.9: ‘Others will talk nonsense, but here we will not talk nonsense.’
I’m sorry
Please, forgive me
I love you
And I thank you
This podcast interview of a female Buddhist Teacher and AI Expert Dr. Nikki Mirghafori of UC Berkeley gives a more nuanced view of AI, making an important distinction between Artificial Narrow Intelligence (ANI) and Artificial General Intelligence (AGI) which operates more human like ways. She ends the interview with the optimistic possibility and potential of Benevolent AI in cleaning environment, curing diseases, improving social well-being and such.
The outcome depends on the intentions and choices of both the users and developers. I feel vigilance and awareness is critical in times of rapid change and transition like now.
Sadhu!
