ChatGPT INTEGRATION PLEASE🔥🔥🔥

That would sure be cool. Imagine being able to ask it questions in fluid English (not having to know some arcane search syntax) and get back answers like this. Then again, seminary professors, beware! This will definitely be used by students to avoid learning for themselves. We need to be sure we are not producing a next generation of seminary grads who can only parrot what AI tells them.

Yes please!! 

Logos Devs - there's a way to "fine tune" ChatGPT with specific content (ex. Content purchased on Logos). I think people would pay well for an ability to chat with ChatGPT that has been fine tuned to Logos' content.

https://beta.openai.com/docs/api-reference/fine-tunes/create

So we can expect something in the beta for v24 then? Thanks Mark!! 😜

I'm sure there are also ethical concerns with how people like preachers and students would be tempted to use AI for their writing. A lot of people's sermons already sound like they could have been just as easily written by chatbots (smh). And we already have proven instances of preachers and scholars plagiarizing. Seems like AI is about to make things even worse.

But I'm all for a customized AI experience for Logos at what ever point that becomes technologically and financially feasible. The cheaters are going to find a way to cheat regardless. Faithlife just has to protect itself from liability for any user wrongdoing. Expect updated terms and conditions whenever Faithlife implements AI lol.

I am not convinced such technology is a good idea in a product Logos, it all comes down to how it was implemented.  I want Logos to be able to better point me to resources that are more relevant to what I am studying, but I don‘t want AI replacing the development of my own learning, reading and critical thinking skills. As it is today most people rely on translations and secondary sources as we don’t know the languages the primary sources were written. To then have AI to replace our need to even read the translations and secondary sources and eliminate the joy of discovery and need for us to think for ourselves, I’m not in favour of such a direction for Logos Bible Software.

I don’t believe this is the intent of Brandon‘s suggestion just putting out my 2 cents.

Kiyah said:

I would be in favor of them using AI to make Faithlife Assistant work better, such as finding information using the resources in my library to answer questions, changing settings, executing commands, suggesting books and articles from my library, or even teaching me how to use a feature in Logos. It would help with accessibility as well. But they shouldn't implement it in a way that it writes sermons and papers for you, which people are demo-ing in YouTube videos right now.

The whole reason we have Bible Software in the first place is to make the work of ministry easier and faster. Physically searching for a word in BDAG while exegeting a passage takes 2 minutes, but it only takes a few seconds in Logos. Physically searching my college library while researching for an essay can take hours, with Logos it is minutes. If AI implementation into logos continues to improve upon the speed at which I can access or process data then I'm all for it. Time saved here behind the scenes allows me more time to speak with people, share the gospel, pray etc.

But getting AI to write sermons or essays for me which I pass off as my own.? Yeah nah that's a hard pass. The work of ministry is the application of the gospel to the lives and hearts of people - I can do that because I know my people, an AI doesn't.

Fair enough, and well done for addressing this directly. Cost wise you could alway offer on the side a subscription based model (faithlife assistant premium 😁)... Like any other technology innovation prices start high but then tends to drop massively, if the only thin holding this back would be coat I'm sure some of us would be willing to front that cost (within reason).

I'm not going to pretend that I know enough about the other issues raised, but makes sense seeing that this technology is still in its infancy. Again I guess it's a matter of time for those issues to resolved/polished, and am glad that you guys have it on the radar. 

Logos 11 maybe?😁😁😁  "Bible study taken to the AI-xtreme with Logos 11" (might or might not have been created by chat GPT)

Rosie Perera said:

That would sure be cool. Imagine being able to ask it questions in fluid English (not having to know some arcane search syntax) and get back answers like this. Then again, seminary professors, beware! This will definitely be used by students to avoid learning for themselves. We need to be sure we are not producing a next generation of seminary grads who can only parrot what AI tells them.

I'm with you on this AI search thingy ....  Then we add voice to that and bidda bang!!!   Less typing!!!! Dr. Spock lives again!!!

I see it now..... "Logos... find the answer to who Cains wife was?"       Seriously... I think AI would help Search.    [8-|]

lol.  I had a mini tape recorder that I carried with me also.... and then when I got to the office....I gave it to my secretary.... she was like a digital helper!  "Julie... get me a red eye to IAH for the 29th coming back on the 8th."  and bingo.... it was done!

I think it will happen with digital things. It seems logical that it's the next big step. I do see a cavity in all this.... that of making bible study mundane and thus people doing less of it.... I don't mind helps.... but nothing can take the place of human study.   And as to God's word.... I think people are dumb enough already... OMHO [8-|]

Doc B said:

Rosie Perera said:

and get back answers like this

So where can I find a teenager (or the equivalent) to show me how to install and set up this thing on my computer? It looks like a heck of a lot of fun.

You don't need to install anything. You go to https://chat.openai.com/chat and set up an account. Then just type your question into the box at the bottom of the screen.

Occasionally you might get a message saying ChatGPT is overloaded and you'll have to try again some other time. But they've been working at scaling it up to be able to handle the higher load with all the recent interest in it. Also occasionally you might have to refresh the browser tab (F5 on Windows) if you've been inactive for a while and come back to try another chat.

All, in all, it's been very educational playing with it for the past couple of weeks. I'm finding what its limitations are (and there are many). But I'm also impressed at how much it can do.

There are all kinds of interesting ethical discussions taking place about this new technology. I wanted to understand it so that I could engage thoughtfully, in a theologically informed way, in such discussions.

I'll leave you with this intriguing article from the Jewish Chronicle:

New York rabbi delivers sermon written by artificial intelligence

And this one in Church Today:

Is Artificial Intelligence the Solution For Helping Church Staff?

The technology behind ChatGPT can be used to enhance a smaller corpus of specialized resources such as those in the Logos library in several ways:

1. Text completion: By training an AI chatbot on a smaller corpus of specialized resources, it can complete or predict missing text in the corpus, making it more comprehensive. This can also help with searches as some have already mentioned in this thread.

2. Text summarization: An AI chatbot can be trained to summarize the information in a smaller corpus of specialized resources, making it more accessible and easier to navigate. For example, ask the Chatbot to summarize an entire journal article in one paragraph to see if it worth reading in more detail (like providing an abstract where one does not already exist).

3. Text classification: An AI chatbot can be trained to classify text in a specialized corpus, making it easier to search and retrieve specific information. It can also tie texts from many different types of resources to make connections that are not readily apparent.

4. Text-to-Q&A: An AI chatbot can be fine-tuned to provide answers to questions posed by users, making the specialized resources more interactive and user-friendly (which some in this thread have also mentioned).

The quality of the output will depend on the size and quality of the specialized corpus used for training (entire Logos library?). The more specific, diverse, and high-quality the corpus is, the better the results will be.

The amount of time it takes to train an AI chatbot on a specific corpus can vary depending on several factors:

1. Size of the corpus: The larger the corpus, the more data the chatbot has to learn from, and the longer it will take to train.

2. Quality of the corpus: The quality of the corpus can affect the training time. A high-quality corpus, with well-written, diverse, and relevant data, will require less time to train than a low-quality corpus.

3. Hardware: The hardware used for training will also affect the training time. Training on a powerful supercomputer will be faster than training on a smaller CPU. There are CONSIDERABLE hardware resources currently required on the backend to make these work well. For example, ChatGPT was trained on an Azure AI supercomputing infrastructure. Somebody has to pay for that computing time...

4. Number of epochs: The number of times the model is trained on the corpus, also known as epochs, can affect the training time. The more epochs, the longer the training will take.

5. Fine-tuning or pre-training: If you are fine-tuning a pre-trained model on a specific corpus, it will take less time than training a model from scratch on the same corpus. This is helpful, for example, when adding new resources to a library that has already been used to train an AI chatbot.

In general, it can take a considerable amount of time to train an AI chatbot on a specific corpus, depending on the size and quality of the corpus, the hardware used and the parameters set. Attention-based models combined with reinforcement learning and other supervised training are usually also required to fix biased behavior and biased results (among other limitations).

All of that said, I can envision a chatbot like interface to enhance Logos that could be very useful for some applications.

P.S. Even though a have a background in AI, I used ChatGPT to help produce this post (with my own modifications) to illustrate its abilities for those who have not yet used it.

Hi everyone - This is a fascinating thread raising a number of issues, I see most respondents appear to favour the integration of ChatGPT or some other AI facility within Logos in the future. 

While I have no difficulty in theory with an AI system being available within Logos - if it were included I would like the facility to switch it off i.e., to deactivate it completely and/or to activate it subject to predetermined security restrictions (eg., to selectively prevent it accessing the internet or any remote server; prevent it accessing my location, photos, documents on my computer etc). As far as possible, I would also want to prevent Faith life or any third party being able to track my use of the AI or retain any records concerning such use. 

Currently, I prevent the operation at any time of programs such as Siri, Cortana etc active on my computers, phones or other electronic devices. For me, such programs present potential security and privacy problems. Its possible these potential problems would be greater for a more sophisticated AI program. 

To summarise - If an AI facility was introduced to Logos Software without security safeguards and user-defined limitations, the Logos program itself could become a serious security problem for me and many other Logos users. Its relevant that we live in a time of growing hostility to Christianity and Judaism in many Western countries, For that reason, we need to be assured of the security implications of any major changes to the Logos program.  Keep well Paul 

My concerns aren't so much theological bias as much as worldview bias that leads to censorship and the ever broadening definitions of concepts such as 'hate speech', which tend to include more of the core teachings of Christianity as time passes. Even now there are topics and subjects that ChatGPT refuses to discuss. Over time this uncertainty could become a real issue if an outside technology such as this becomes an integral part of Logos. That's why I believe it important that such a product only be included in Logos if such scenarios can be reasonably mitigated by Faithlife's own implementation that they have a reasonable level of control over.

MJ. Smith said:

Paul said:

If an AI facility was introduced to Logos Software

I think you have too narrow a definition of AI ... most language processing has elements of AI especially translation processing such as the DeepL and Google Translate that were added in L10. I suspect there was AI previously but I am not privy to what is going on behind the scenes - it may be that AI was merely used to obtain the data for particular features.

As I'm sure you realise, I was not seeking to define AI itself nor its current reach (in some form) in our existing Logos 10 program. Nor am I particularly curious as to what happens behind the scenes in Faithlife.  My concern is a basic one concerning the possible future introduction of a new AI system (however described) which may raise security and privacy issues. In the excitement many people seem to feel over "new toys", sometimes security is forgotten as an issue. For me and I suspect many others, maintaining the security of our computers and database is paramount.   

I used ChatGPT to help me develop a sermon series planning calendar. It did a fairly good job. I see how it can help in project and planning development. I suspect it can be customized to search within the Logos environment and make search more efficient. The Faithlife Assistant probably will see an improvement with AI integration. 

If anyone here is a chess nerd, ChatGPT is quite hilarious playing the royal game. (Warning: Nerd humor is, well, nerd humor.) Check (bad pun) this out: https://www.youtube.com/watch?v=rSCNW1OCk_M 

It wasn't quite that bad when I played it, avoiding illegal moves until move 21. I mated it on move 22. It's interesting, IMO, to see how it responds to corrections. It's also interesting (and often, wildly irrelevant) to read the descriptions about why it makes certain moves.

I had similar success asking about music asking which songs have similar chord progressions to another song I mentioned. It's often radically off but weirdly succumbs to my gentle corrections. From a conversational standpoint, it's pretty good at making grammatical sense, but pretty bad at making reasonable statements, and much worse at creating extended expositions of information.

I'm not quite ready to turn my theology over to it. [:O]

I love this idea and totally agree. This would be invaluable. I have so many resources available to me in Logos that there is no easy way to reliably survey them all. But an AI assistant could help synthesize things!

MJ. Smith said:

I appear to be in the extreme minority -- I have not been impressed by ChatGPT

AI is very new and exciting to people. I think people on this thread are thinking of its possibilities.  In the next few years, I suppose AI will be much better.  And Faithlife hopefully (in the background) is beginning to think about its potential to enhance or destroy their software.

Mark said:

MJ. Smith said:

I appear to be in the extreme minority -- I have not been impressed by ChatGPT

AI is very new and exciting to people. I think people on this thread are thinking of its possibilities.  In the next few years, I suppose AI will be much better.  And Faithlife hopefully (in the background) is beginning to think about its potential to enhance or destroy their software.

I am also unimpressed, but the future possibilities are intriguing. To revert to the chess comparison, in the late 1960s, the English International Master, David Levy made a bet that no chess computer could beat him within 10 years. There is now no human player that could come close to the playing strength of the AI beast we can put in our pocket. It will be interesting to see what develops over the next couple decades, but I am much more comfortable understanding and negotiating when I understand another person's opinion, and how their (and my) self-interest lies behind their (and my) judgments in theology or whatever. While chess is technically definable in completely objective terms (although perhaps an SSD the size of the universe could not contain a database of all possible games), it is possible that human interactions with God -- and indeed, how to talk about human interactions with God -- are not so mathematical.

But concepts like "convincing arguments" could be statistically compared, so is it conceivable that an AI program similar to Alpha Zero could be tuned based on human responses to their arguments? And doesn't this open up stuff about the discussion of the Holy Spirit's part in wooing human hearts based on the words that someone says? If nothing else, it may force us to refine our language when we talk about passages like 1 Thes 1:5, 1 Cor 2:1, and 2 Tim 1:7.

It all seems weird to think about now and maybe scoff at such a notion, but I already regularly phrase emails written on my phone using the suggested phrase because it's easier than typing out precisely what I would've said otherwise.

And yet, I can see huge value in all of this. But like any technological advancement, I hope we can be discerning in its use, something humanity has not demonstrated in the past.