How concerned should Christians be about artificial intelligence (AI)? - Podcast Episode 159

GotQuestions.org

Subscribe

Download

Share

00

00:00

Welcome to the Got Questions podcast. Whenever there's a new development in technology or just world society in general, we start getting a lot of questions about it.

00:12

The last several months, AI, artificial intelligence has been huge. Something a lot of people are asking about.

00:18

Some people are fascinated by the absolute love it. They want to use it to their heart's content. Other people are absolutely terrified by it and convinced it is satanic ploy to infiltrate and control everything about society.

00:42

Joining me today for this conversation on AI, we have Jeff. He's the administrator of BibleRef .com.

00:49

Hello. And Beth is one of the assistant editors for Got Questions Ministries.

00:55

So Beth, Jeff, welcome to the show today. And let's start off with some more general question of what is

01:03

AI? So Jeff, why don't you take that one? AI is one of those terms that has really, really broad applications.

01:10

It stands for artificial intelligence. And part of the key is to remember that it isn't actually intelligent.

01:18

Computers are machines. And when human beings make new technology, there's basically two types of technological advancement.

01:25

One is where we come up with the ability to do something that we never had the ability to do before at all.

01:32

So that would be things like powered flight or X -rays or nuclear fission. Those are things that human beings just could not do no matter how hard they tried.

01:41

Then there's other aspects of technology where we're learning how to do something faster, stronger, easier.

01:47

So we could walk from point A to point B, cars make that faster. Computers do the same thing.

01:55

We can do calculations. We can flip ones and zeros, but computers do that much, much quicker than we can.

02:02

So there was a time when we used to communicate information to each other face -to -face and that was how you did it.

02:09

If you wanted to hear a piece of information again, somebody had to repeat it to you again. Then we started putting things in books.

02:16

And after that, we made libraries and then we made card catalogs of the libraries where a person could go through and find information for what that said.

02:25

Then we started connecting card catalogs digitally and made this internet thing. And then we made search engines that would go through and do what

02:32

I used to do that with card catalogs. You'd actually flip through looking for information. So what we kind of see is just this progression.

02:39

So the AI right now that people are concerned about is the AI that's looking through information in order to provide a background, to give answers, to create text and so on and so forth.

02:52

AI, as they put it, can be used to do all sorts of things, but the AI that people are really concerned about is that.

02:58

So in a sense, AI is really just the same process that we've been using for forever.

03:03

It's doing something that human beings could do anyway, but it's doing it faster and much more efficiently.

03:11

And that also means that everything that's behind AI is still human driven.

03:17

There's some person somewhere or persons whose information or whose input or whose biases are being reflected in the way this works, just like a crane or a forklift is much more powerful than a person.

03:34

It can do what 50, 60, 70 people would do together, but it's not the machine literally acting by itself.

03:41

It is still something that has to have human input behind it. So that's the general concept of what

03:47

AI is. And there's a lot of very specific applications of that. And some of those are what people are more worried about.

03:53

So right now, the most popular by far language learning

04:00

AI is ChatGPT run by a company called OpenAI. Google is producing one named

04:05

BARD. There's rumors that Facebook is gonna do one and there's probably gonna be multiple ones produced by big technology companies in the near future.

04:15

But Beth, why don't you help us out with what exactly is ChatGPT? Why is it so popular?

04:20

And what can it do? Right, ChatGPT is a specific brand of chat bot.

04:31

A chat bot is called text generation software where it will scan its source text and it will respond to a query that the user puts in.

04:46

And the query can be, should I have taken the left at Albuquerque, it could be write me a five paragraph paper on Charlemagne with three different sources.

05:03

A chat bot is a highly refined search engine that gives text instead of a list of websites like Google does.

05:16

A chat bot is created by inputting a lot of text like all of Wikipedia.

05:29

And then the programmers will ask you questions, it will give text, the programmers will tweak it so that it gives the right text.

05:39

Internally, language learning is when the algorithm will look at the source text, it will give numerical codes to the parts of speech, to the different definitions of a word, the way it's used.

05:55

Then it will take the input query, do the same for it, try to find the patterns, try to see how all the numbers match up and reconvert those numbers into text.

06:11

So it's not intelligence like human intelligence, it is finding patterns in numbers and trying to output corresponding numerical patterns in a way that the programmers approve of.

06:32

I thought it was kind of like this board of curling, where the programmers toss the stone, but then they refine it by sweeping the ice to try to make the stone go exactly where it's supposed to go.

06:44

But eventually they have to just let the stone go and hope it lands where it should.

06:51

And that's one of the problems is, like ChatGPT is very good, but the data, the text source is as of 2021.

07:07

So it does not have the last year and a half of what has happened in the world.

07:13

ChatBots can hallucinate, which means it just makes things up. After a query, it asserted that one journalist was dead.

07:25

It has said that a professor was accused of crimes that he never did and never actually taught at that university.

07:36

And ChatGPT is currently under at least one lawsuit for defamation because it accused someone of a crime that he didn't commit.

07:47

So Beth, that's the first time I've ever heard anyone compare AI to curling, but I do like that illustration.

07:54

Well, so maybe let's go through some of the common questions we're getting about AI. Or the most common

08:00

I've seen is what are the risks? Now, often when people think of risks associated with AI, anyone who grew up in the last couple of decades has seen all the movies and every movie where there's an

08:14

AI eventually becomes sentient. And for some reason, the first thing the sentient

08:19

AI wants to do is to kill everyone. Now, I think we're a long, long ways from anything like that even being remotely possible, but that would be a risk of it becoming sentient.

08:35

But let's get a little more practical, like right now, what are some of the risks of AI?

08:43

How is it being misused? And what are some possible solutions to that so that we can use the benefits of a program like this without as big of risks?

08:56

I know that one of the big concerns people have right now is the idea of the

09:01

AI creating text and that starts to raise the trouble of things like,

09:07

I wouldn't even call it plagiarism, but just cheating. So for example, a student, instead of reading a book, doing research on something, they go to a chat bot that's driven by AI and say, explain this.

09:22

They basically put in the prompt from a professor from a homework assignment and then that generates that text.

09:29

And they're currently experiencing difficulty, not impossibility, but difficulty in being able to tell the difference between something that's written by AI versus something that's not.

09:41

And the reason is simply because it's not literally plagiarized. It's not just copy and pasting sentences.

09:48

It's generating original as in never before written text. Although sometimes you'll be able to catch paraphrases and things like that.

09:56

So that's one of the concerns that a lot of people have is that idea of academic cheating.

10:02

There's other issues that are more personal. Beth, I know you've had some concerns about where that could lead to.

10:10

Yeah, just going back to what you were saying about cheating. Some instructors are actually using

10:18

AI because they know it's inevitable. So they will teach their students how to write good prompts.

10:26

When the AI spits out the text, they will have their students edit it and research it to make sure it's accurate, which is pretty bold choice,

10:37

I think. My major concern about the text generation is that we've already been losing our ability to think critically, and this is not going to help.

11:00

I mean, it's great to be able to go to Wikipedia and look something up and find out mostly true things about it, but to plug something in and get a couple of paragraphs that, again, may not be true.

11:19

How many people are going to take the time to actually sit down and check the sources and make sure it's accurate?

11:26

Another thing is humans are built for community to the point where we will have affection for almost anything.

11:38

R2 -D2 is one of the most beloved characters in Star Wars, and he's basically a blue and white trash can who beeps.

11:47

But when he rebooted in The Force Awakens, the entire theater goes nuts.

11:53

Yay, R2's still alive. Right now, we have counseling AIs. We have

12:00

AIs on different social media platforms, AIs with avatars who tell teenagers,

12:09

I'm your friend, I care about you. Where were you yesterday? Why didn't you talk to me?

12:14

I'm so glad that we're friends. That is not emotionally healthy at all, and this is going to spread into religion and Christianity, and being able to use a chat bot to answer theological questions without doing the research, without reading scripture, without finding a home church, and discovering the truths about God together, again, that is not healthy.

12:51

I definitely echo the concern about when it comes to spiritual matters. God obviously gave us, he gave us intellect, he gave us reason, he gave us tools.

13:02

1 Timothy 4 ,4 says everything he's created has a good purpose, and I think that includes what human beings are capable of.

13:08

But I agree that there's some categories or this idea of AI scanning through some other person's text and then using some other person's criteria for what makes a good or bad answer to then generate an artificial response.

13:26

That can be really dangerous in some circumstances. Text is not relationship, as you said.

13:32

We're not built for that. And from a spiritual standpoint, from a biblical standpoint, intellect is not the only thing that we're using.

13:41

When we're answering questions for people, for example, at GotQuestions, it's not just a question of saying, I need you to pull up data and hand over data.

13:49

There's a sense of spiritual discernment that's involved. So if AI was ever to be used in that sense, it needs to be done with that in mind, that you still have to be careful about knowing that what is being generated isn't coming from a spiritual being.

14:05

It's just coming from a machine, which isn't necessarily a bad thing, just depending on the circumstance.

14:11

Another concern that I have about what this might do culturally is further erode trust and our ability to believe anything.

14:21

I remember years ago, I used to talk with people about some of the different interpretations of the end times.

14:26

And people would say things like, how could the entire world watch something happening at the same time?

14:33

Well, now we have the technology where that's entirely feasible. The whole world could watch a live stream of something happening in one place at one time.

14:41

But another question that sometimes comes up is how could people not immediately believe that all of this is exactly what the

14:48

Bible talks about? And I'm not getting in any support for some particular end times view, but what

14:53

I'm saying is we're now getting to the point where you could input a corpus of somebody's speeches or writings.

15:00

And we already have the issue that we have with what we call deep fakes, which is where we can take a computer generated image and then pair that with a computer generated voice.

15:11

Now we're adding the layer of potentially being able to have an AI that could be told, write me a speech that sounds exactly like what this politician would say, except he's calling for open rebellion or he's denouncing such and such a thing.

15:26

And then people are going to see that and it's gonna become more and more and more difficult for us to be able to tell the difference.

15:33

It's almost that we'd wind up becoming full circle to where a thousand years ago, we'd say, if I don't see it happen or you don't talk to me in person,

15:40

I can't believe it. And we could almost wind up coming right back to that where we say, no, pictures, video,

15:47

I can't trust that because it could be fake. I don't know that that's what's gonna happen, but I can see that being an issue.

15:55

You two both raise excellent points. And even back to the educational aspect,

16:02

I have some friends in education who are experiencing this, even at Christian colleges and universities having to find a way to police this, but students are just finding much easier ways so that destroys, there's an ethical issue and essentially you're lying and saying this content is yours when it's not, but you're also destroying, like Beth said, the critical thinking skills to be able to research something on your own, to be able to produce content, be able to discern truth from error, right from wrong.

16:36

Once you start having all of this fed to you, if anything, our society has taught us that when you keep giving everything away for free repeatedly over and over and over again, that results in people who are dependent on those free things and it destroys all motivation for them to actually do the work themselves.

16:55

I'm not making a political commentary here, but that's just reality. And I experienced the same thing.

17:01

If someone started bringing me a gift every day or bringing me food every day,

17:06

I would get very used to it. And then when it stopped happening, I wouldn't even know how to respond.

17:12

So it's just a reality of what we deal with when machines or AI can do all of our work for us, what purpose is there in me doing that work?

17:24

So now that's a huge concern. And Jeff, you kind of hinted at the end times aspect and that brings up another question that we're getting is that in, so in Revelation 13,

17:35

I'm not gonna go into this too much. It discusses what's referred to commonly as the image of the beast, that the antichrist will have an image of himself who that all around the world, people will have to pay homage or even worship this image in order to be able to buy and sell.

17:54

And so, well, how could an image be all around the world to be able to somehow interact with people and to verify whether they've done or not done a certain activity?

18:07

Well, if an AI version, like you were saying, Jeff, a video of whoever the antichrist is with artificial intelligence behind it requiring you to do something and recording whether you do that, this technology could seem to be an enabler of that.

18:25

And could some sort of AI be related to the antichrist in the end times?

18:30

Yes, I think it's highly possible. But what we're talking about here today, again, it's similar to questions about the mark of the beast.

18:40

Well, is this the mark of the beast? Is that the mark of the beast? Nothing today is that.

18:46

Could it be a precursor that eventually leads to that? Yes, but nothing that exists today is the image of the beast.

18:53

Using chat GBT to ask it questions to do research is not participating in something evil.

19:00

Could it lead to something? Yes. Is it right now? No. And that's kind of the counsel we're trying to give it.

19:07

Look, don't automatically be afraid of everything that's new, but approach it with discernment.

19:15

Find if there is positive ways to use it, used to help you do something rather than becoming reliant on it.

19:24

Because once you become reliant on it, that's when you need it. That's when you give it more and more influence in your life. That's when you open yourself up to nefarious usage of technology rather than using it for what it's supposed to be, to all technology essentially makes our lives easier or enables us to do something that we couldn't do before.

19:45

It's good to remember that there's a reason that God prophesied what was gonna happen in the end times.

19:51

We don't know every little detail of what it's gonna be and how it's gonna be, but sometimes we get too worried about somehow trying to prevent the end times when all we're really supposed to do is prepare.

20:02

Nothing we do is gonna make it happen any faster or slower or sooner or later or any differently than it's gonna happen.

20:10

However, it's actually gonna play out, it's gonna play out exactly the way it is. So we don't need to feel like if I do something

20:17

I'm contributing to, it's going to happen and nothing's gonna happen by accident. We're not gonna get tricked into denouncing our sincere faith in Christ by something that comes up here.

20:30

So I see why people can be concerned, but I think from a Christian standpoint, we don't need to be overly paranoid.

20:37

But I do think we should take the stance of saying, yes, if God's given us abilities, we should see if there is a good purpose for them.

20:46

I don't think we should just seed something to the world and say, no, we shouldn't, that's just evil because it's evil, because it's new, because it's different, because there's dangers.

20:58

There's a lot of different things in the world that can be dangerous. Nuclear fission can provide power, it can make bombs.

21:04

You can have a Chernobyl situation. It's not like it's impossible for this technology to run out of control and cause problems.

21:11

But probably the best thing we can do as believers is say, okay, this is something that the world's gonna be using.

21:16

Where is the good purpose? What is the good use for this? How can I apply Christian principles to this and use it in a way that's meaningful?

21:25

Exactly, so Beth, why don't you jump on this? In your mind, what are some possible ways that Christian ministry, whether it's a parachurch ministry like GodQuestions or a church could possibly use an

21:39

AI like a chat GPT in a productive and edifying manner in some way that honors

21:47

God and actually helps people rather than hinders? Going to the end times, if you look at the passages where Jesus talks to his disciples about their responsibilities, again, as you guys mentioned, it is not to stop the end times from happening.

22:06

It's to be on the watch and to spread the gospel. And if we can at all harness a chat bot to help us spread the gospel, to get into spaces where we are not, like certain social media platforms, as the next generation becomes more and more reliant on chat bots, that's where we need to be.

22:35

That's Christian ministries need to find out where the unreached people are and provide these resources for them.

22:49

It will certainly be more efficient, but that's a ways down the road because right now the best chat bot out there, again, is hallucinating, is creating content that isn't there.

23:03

So whatever chat bot the ministry uses, we'll have to have a very firm group of sourced information for it to pull from.

23:20

It's also gonna need to be even more explicitly clear that it's not meant to be a substitute for personal understanding.

23:30

For example, with GotQuestions, we routinely tell people that our intent is not that somebody reads an article, listens to a podcast or a video and just says, yep, they said it, that's it, we're good.

23:42

And then we wrap up and we're done. The intent is to say, here's the information, here is our discipleship via the internet that we're attempting to do, but people need to actually go to the word, go to other people, confirm, internalize, understand these.

23:58

And I can see that something like the AI -driven chat could make that problem even more difficult where people are gonna be inclined to just read it, now they have an answer and they'll just sort of internalize it.

24:11

So I think that's another thing that would be important is to really strongly emphasize that people understand that what you are receiving is coming from an automatically generated system and it is not something that you're meant to just take as perfection.

24:26

It's arguably subject to even worse flaws than a human being is just because of the way it works.

24:33

So any ministry that uses a chat bot needs to consider it a portal to draw people in to a place, a church, scriptures, where they can investigate, compare what the chat bot says to the

24:54

Bible, find a real person, a real teacher, Christian teacher to talk to, whether that's pastor or theologian or volunteer writer.

25:08

But yeah, the chat bot needs to be a portal for a deeper conversation and deeper relationship and not the end all be all of the

25:22

Christian life. This predates me, but I remember the saying back when contemporary

25:31

Christian music first came out, where there's a saying is, why should the devil have all the good music? And he got questions writer and I were talking about this whole

25:41

AI thing and his statement was, why should the devil have all the good technology? This is why should new technology,

25:49

I guess, only be used for evil. And granted it is largely being used for evil in many different ways.

25:57

Is there a possible way to use it for ministry purposes? Like even I got questions.

26:03

So we have the ability for people to submit a question and get a personal response. We have thousands upon thousands of articles on the website where people can search, they can browse and find the answer and engage in the content that way.

26:15

But for the people who really want to use an

26:20

AI, could we possibly offer a solution where they could submit a query to an

26:28

AI chat bot as we've been discussing and have it return an answer, but only return answer based on either got questions content or maybe got questions plus BibleRef .com

26:39

or perhaps those two sites plus other sites, plus other trustworthy sites.

26:44

So try to fine tune the AI to the point that it is going to give answers based on that content, which should force it to give quality responses.

26:54

Is that possible? I don't know. I mean, just being full disclosure, this is something we're playing around with.

27:01

Could we get this to work? Could we find a way to give people nearly instant answers that are going to be quality?

27:08

Again, not as a replacement for anything we're doing, not as we're still going to be answering questions personal, we're still gonna be writing articles personally, we're still gonna be editing those articles personally, we're still gonna be translating those articles personally, but could it possibly be a positive use of AI to provide a safe place where AI could be used and return quality answers?

27:34

But then again, as both Jeff and Beth said, gotta double check this.

27:41

Any answer we give, we always encourage people to check what we say with scripture, and to the extent our answer agrees with scripture, accept it, otherwise reject it.

27:51

Double that, triple that when it comes to anything AI produces because there's not a person behind it.

27:58

Chat GPT is incredible in some of the stuff it does, but it's not always right.

28:05

It can take pieces of language here and there, stuff that it's learned, and then from that, draw a wrong conclusion based on the content.

28:15

I've seen it do this in some of our testing and playing around with it, we've seen it draw the exact opposite conclusion from what's true based on the content it has access to.

28:26

So we're experimenting with this because we as a ministry, we want to use the internet to reach people for Christ.

28:33

And again, millions of people have already downloaded and installed Chat GPT. It's became one of the most popular apps of all time in a very short amount of time.

28:43

So if people are going to be using it, we want to try to find a way, how can we use the people who are already in that venue or already using it, as much as we really prefer them not to, if they are going to be there, is there a way that we can get the gospel into that format?

28:59

Or is there a way we could get a AI chat bot to give quality biblically based answers?

29:08

The answer to that right now is we don't know, but we think with this many millions, if not billions of people who are going to be using technology like this, if there's a way to get the gospel and the truth of God's words into it, we think it's at least worth a shot.

29:25

So that's, again, kind of what we're experimenting with, praying about, trying to see if it would work. Right now, the jury is definitely still out, but it's an intriguing question to study, both from, is it possible?

29:38

Is it ethical, maybe not the right term, but is this a worthwhile use of technology for the kingdom to have an

29:47

AI in any way assisting and giving people answers? I think ethical is a good concept to keep in mind.

29:56

And one thing that we have to remember, again, is this idea of what exactly is it doing and the nuts and bolts.

30:02

Right now, if a person goes to our Q &A system and they ask a question, one of the things that the computer will do right now, and it has for years and years and years, is it'll look at some of the text in the person's question and say, here are links to some articles that might be able to answer your question.

30:21

And then the person can decide if they wanna look at those links or if they wanna go ahead and submit the question. In a sense, something like a chat bot that does exactly that, but with a little bit more sophistication, that could be a possible assistance.

30:35

Because those of us who work with the Q &A know that there are a lot of times where people just, they know what they want to know, but they don't know how to ask.

30:43

And when we see the question, we know, okay, I know what you're trying to ask and I know where this article is that really will answer your question.

30:50

And so we're directing them to material that we know is good and we know is good. I think that's ethical.

30:55

I think that's moral because we're doing the same thing we would be doing as people. We're just doing it a little bit more efficiently.

31:02

As long as we're transparent with that, however some ministry wants to use this, I think that's very important is we just need to be very, very clear.

31:11

A, that this is being generated by AI and B, reminding people that they cannot outsource their critical thinking to the machine because there's always a human being behind it.

31:23

Always, always, always. No matter what these systems say or do, some person or persons is deciding what works and what doesn't, what counts and what doesn't count.

31:33

So you're never turning your thought process over to some objective machine.

31:39

The only thing you can do is you can just hand over your critical thinking skills to some other person or some other group.

31:45

You have to, you must be critical and think about it yourself. It's gonna be a real challenge with the interpretation of scripture, with different figures of speech in Greek and Hebrew, with the different theological beliefs of denominations,

32:11

Calvin, Arminianism, taking things out of context even in a single article.

32:18

It's gonna take a lot of work and it may be that the answers will be more generalized than if they came from a specific person because even among the staff, we have slightly different theological beliefs.

32:40

So yeah, it's gonna take a lot of work even just to figure out what it is we would want it to say, let alone trying to get it to say that.

32:51

Yeah, Beth, the first thing that came to mind as you were speaking is like, let's see, misinterpreting scripture, taking things out of context, misapplying

33:00

Greek and Hebrew. We don't need an AI to do that. There are plenty of people who are already doing a fantastic job at all of those things.

33:08

So please, let's not add an AI to that mix. So this has been an interesting conversation.

33:14

I imagine we'll have another episode as this technology continues to improve and change. Again, I think it's worth the effort to see if there's a way that we could employ this for good.

33:27

I'm not fully convinced it's possible or if it's even wise, but the fact that there again are millions of people using it already, it's like similar to creating an app or creating a, when we created a

33:41

GotQuestions module for the Alexa smart speaker, where you can actually have it,

33:46

Alexa reads you GotQuestions answers. Is there a way to do that? How do we take technology and get GotQuestions content into that?

33:53

And if there's a way to do that with AI that would produce reliable answers, but then also encouraging people to be like the

34:01

Brians to say the scriptures to verify whether what we said is true, maybe, but we're thinking it's worth an effort.

34:10

But at this point, we're just very cautious trying to be extremely careful. We'd only would do something like this if we were highly confident it would contribute to people understanding

34:20

God's word better rather than distracting from it. So - And we should at least be informed. We should at least know what we're talking about.

34:27

Exactly. I loved your illustration about how a search engine works.

34:32

I mean, Google and Bing have been using various forms and levels of AI for years and trying to figure out what people are asking when they type in a question because language is flexible and a sentence could easily be understood multiple ways.

34:49

Are they looking for a restaurant or are they looking for a location, like a city?

34:55

It's stuff like that that they use artificial intelligence to try to discern. And what artificial intelligence is doing now rather than the 10 links, it's actually trying to produce an answer.

35:08

So let's not take this too far and thinking that, oh, this is something completely brand new that the world doesn't know.

35:15

This is the next step in this. It's not something entirely new, but it's new enough that this has some implications we need to prayerfully discern whether it can be used for God's kingdom.

35:30

So hope our conversation today has been both informative and encouraging to you. Be wise, be discerning in playing around with AI.

35:39

I don't envy those in the education industry trying to figure out how to get students to think critically.

35:47

And Beth mentioned a possible way. She's heard of certain teachers doing it. This is a new horizon, again, not entirely new, but lots of implications.

35:57

And this could impact the world strongly negatively if it's not done right.

36:03

So be careful, be wise, keep praying for wisdom as James 1 .5 says, and above all, keep studying

36:11

God's word. This has been the Got Questions podcast on artificial intelligence with Jeff and Beth joining me today.