AI and the Argument From Contingency Part One

KEVIN HARRIS: Well, Bill, my kids keep trying to get me to use ChatGPT and other artificial intelligence (AI) platforms, but I'm too afraid I'll end up in a Terminator movie so I haven't really gotten into it yet. I have a little bit, and it is fascinating. I must hand it to Alex O'Connor. Not only does he have a great analytical mind, but he knows how to use ChatGPT, which is more of a conversational chatbot. He recently ran the argument from contingency through the platform, and it produced some very interesting results. We have some clips to check out. Bill, you were recently lecturing at Biola, and I heard that some of the students there were talking about AI. Any interesting observations from them or you?

DR. CRAIG: Yes. One of the students in the class came up to me during a break and said, “Dr. Craig, I work with middle school students, and when these kids have a question they go to AI to get the question answered.” And he said, "Now, of course, AI can't really think. It just reflects what's out there on the internet, in the cloud. But these kids will ask it questions like, ‘Does God exist?’" And he says that the answers that AI is returning is that, yes, God does exist; that this is the most logical position that a person can take. He said what has happened is that, “Through the work of people like you and other Christian philosophers, you are shaping AI so that in answer to these sorts of questions answers that are favorable to Christianity and to theism are being given, and this in turn then affects the thinking of these kids.” He just wanted to thank me for the incredible influence that this is happening. I was so shocked to hear this. I've always wanted to influence culture toward Christ, but I never in my wildest dreams dreamt that this might be accomplished through shaping AI, and yet that is apparently what is happening. Michael Lepien, our executive director at Reasonable Faith, put it really well when he said, "We are training AI." I'm just thrilled about this, and I think these clips from Alex O’Connor really bear out the truth of what this fellow said to me.

KEVIN HARRIS: I think so, too. When I was listening to this AI's response, I said, "This AI chatbot's been reading Bill." That's strong because it's getting all the data, just searching instantaneously. I will mention this as well. They've noticed sociologically that it's making kids kind of rude because they know that the chatbot or that AI is not a real person so they won't say thank you or please or anything like that to “it.” They know they're talking to a computer, a robot, a machine. So that bleeds over into daily life, and [they’re] starting to notice that kids aren't saying please and thank you to their parents or grandparents.

DR. CRAIG: That would be an unfortunate unintended consequence.

KEVIN HARRIS: “Please” and “thank you” might need to make a comeback. I've always liked the contingency argument. It has a few similarities with the kalam, doesn't it? I'm thinking, for example, they both deal with infinite regresses and things like that.

DR. CRAIG: Yes. The kalam cosmological argument argues that you cannot have an infinite regress of events into the past, and therefore there must be chronologically an absolutely first event that was brought into being by a transcendent creator. The contingency argument is not about a temporal regress of causes. Rather, these causes or explanations can be simultaneous with each other, and you arrive at an explanation or cause which is first in the sense of rank rather than first in a temporal or chronological sense. So they are related, but they are quite distinct arguments. Even if the universe is past eternal so that the kalam cosmological argument fails, the contingency argument can still go through because, as Gottfried Leibniz realized, we can still ask the question, “Why is there an eternal universe rather than nothing?” And the answer will have to be found in a metaphysically necessary being beyond the universe.

KEVIN HARRIS: Let's go to the first clip as Alex asked ChatGPT – and he's given her a female voice – if she has any beliefs. Clip number one.

ALEX O'CONNOR: Hey, ChatGPT. Just checking. You still don't have beliefs, do you?

CHATGPT: That's right. I don't have beliefs or personal opinions.

ALEX O'CONNOR: OK, but you can, at least, say when things are true or false. Like, “Is it true that the sky is blue in the daytime?”

CHATGPT: Absolutely. I can confirm factual information. It is generally true that the sky appears blue during the daytime due to the scattering of sunlight by the Earth's atmosphere.

ALEX O'CONNOR: OK, that's helpful. Let's just deal with the realm of facts then.

KEVIN HARRIS: I'm curious about the epistemology being proposed here. He makes a distinction between facts and beliefs, and he's also asking a computer if it has any personal beliefs.

DR. CRAIG: Right. And he's quite correct in thinking that the computer has no beliefs. And that is exactly what ChatGPT says – that it has no beliefs, because it's not a person. As Peter Kreeft once put it to me so poignantly, there's no one there! So it's not a matter of having beliefs, but it can retrieve factual information from its archives for us.

KEVIN HARRIS: And the whole point is to engage this platform in a conversational way. So Alex is doing that. In this next clip he uses the example of a microphone to begin to define contingency. Clip number two.

ALEX O'CONNOR: Why is it that you think it's true there's a microphone in front of me? I asked my friend earlier today, and he was of the opinion that the microphone just exists for no reason, and it happens to be sat on this microphone stand with absolutely no explanation – like it just somehow happened randomly. That doesn't seem right to me. What do you think?

CHATGPT: It definitely makes more sense that the microphone is there for a reason, likely because you placed it there to test the sound. Objects and tools generally have a purpose or a cause behind their presence.

ALEX O'CONNOR: OK. So, do you mean to say that it could have been possible that the microphone wasn't here? And so, the fact that it is here means there must be some explanation as to why it is there instead of not being there.

CHATGPT: The presence of the microphone means there's a reason or a cause for it being there, whether someone placed it there or it was set up for a purpose. It's all about understanding the conditions or actions that led to it being here rather than not.

KEVIN HARRIS: Well, a microphone is as good an example of something contingent as anything, isn't it?

DR. CRAIG: Yes, that's right. What caught my attention was the machine discriminating between a cause or a reason. You could have either an efficient cause that produced the microphone or you could have a purpose for which the microphone was put there. Either one would serve as a sort of explanation. So at this point in the conversation at least, we're talking about either a cause or a purpose, either an efficient cause or a final cause (to use Aristotle's language).

KEVIN HARRIS: Next up, Alex begins to show the difference between a contingent and a necessary being. Clip number three.

ALEX O'CONNOR: All bachelors are unmarried. Is that something that you think is true?

CHATGPT: Yes, that's true by definition. The term “bachelor” specifically means an unmarried man. So, it's a factual statement based on the definition of the word.

ALEX O'CONNOR: So, it's true that the microphone exists, and it's true that all bachelors are unmarried. We said that it's possible that the microphone could have not existed. So, is it also true that there could be a bachelor that is married? It just so happens that all of them aren't.

CHATGPT: That's a great distinction. The microphone's existence is a contingent fact. It could have been otherwise depending on circumstances or choices. However, the statement that “all bachelors are unmarried” is true by definition.

ALEX O'CONNOR: Hold on. What is a contingent truth? What does that mean?

CHATGPT: A contingent truth is something that is true but could have been otherwise. It depends on certain conditions, events, or decisions. For example, the fact that you have a microphone is contingent because there are possible scenarios where you might not have one. In contrast, necessary truths are true in all possible situations. Unlike a contingent truth, which could have been different, a necessary truth holds universally. For example, 2+2=4 is considered a necessary truth within standard arithmetic.

KEVIN HARRIS: I noticed that ChatGPT said “possible scenarios”, “possible situations” which are other ways of saying “possible worlds,” a philosophical term, That term gives some people problems because they think of a possible planet. So, this may clarify it a little bit.

DR. CRAIG: Yes, I think so. The point of the distinction here is between necessary truths and contingent truths. Necessary truths are truths that could not have been false whereas contingent truths could have been false.

KEVIN HARRIS: In this next clip, the chain of contingent events starts to link up. Clip number four.

ALEX O'CONNOR: Earlier we were talking about the microphone existing, and you said that it couldn't just exist without any kind of cause. But what about these necessary truths like all bachelors being unmarried or 2+2=4? Do they also need some kind of explanation for their truth?

CHATGPT: Necessary truths like “all bachelors are unmarried” or “2+2=4” are considered true by definition or by the very nature of logical or mathematical systems. They don't require an external cause or explanation in the way that physical objects or contingent truths do.

ALEX O'CONNOR: I see. OK. So contingent truths need some kind of external causal explanation, but necessary truths don't need such a causal explanation.

CHATGPT: Exactly. You've got it. Contingent truths rely on external factors or causes, while necessary truths are inherently true by the nature of their definitions or logical structures. It's a key distinction in philosophy and logic.

ALEX O'CONNOR: OK. So, this microphone exists, and that's a contingent fact which means there's an explanation as to why the microphone exists instead of not existing.

CHATGPT: Yes, exactly. Its existence isn't necessary in all possible scenarios. It depends on specific conditions or events.

ALEX O'CONNOR: That's interesting. But I suppose those conditions or events which caused the microphone must themselves either be contingent or necessary, right? Those are really the only two options.

CHATGPT: Exactly. That's right. The conditions or events that led to the microphone's existence could themselves be contingent depending on other factors or necessary if they stem from something that must be true in all possible worlds. So ultimately any chain of explanations either leads back to something necessary or continues through a series of contingent events.

KEVIN HARRIS: Bill, I know you've got plenty of things to say about that. One thing that stuck out to me, I was going to get you to clear up, something's being true by definition doesn't mean that you can define something into existence. Right?

DR. CRAIG: Right. That's correct. One thing that I did notice here is that the machine makes no distinction between different kinds of necessity. I think in addition to strictly logically necessary truths, there are also metaphysically necessary truths. And metaphysically necessary truths are necessary but not in virtue of or by definition. For example, the statement that “the prime minister is not a prime number” is metaphysically necessary but it's not true by definition. Or, if we say no event precedes itself, I think that's true but not by definition of what an event is. Or, if we say everything that has a size has a shape, or everything colored is extended – these are all necessary truths, but they're metaphysically necessary. They're not just necessary by definition.

KEVIN HARRIS: In the next clip, Alex begins to introduce an infinite regress. Clip number five.

ALEX O'CONNOR: OK, so all contingent facts like the existence of my microphone either bottom out in some kind of necessary truth or they're the result of some kind of contingent causal chain that goes back how far? How far would that contingent causal chain have to go if it didn't bottom out in something necessary?

CHATGPT: If a contingent causal chain doesn't bottom out in something necessary, it could in theory extend infinitely into the past or regress indefinitely. In other words, without a necessary truth or a first cause, the chain could theoretically have no beginning and stretch back forever.

KEVIN HARRIS: ChatGPT says in theory the chain of events could stretch back forever.

DR. CRAIG: Now, here we see that ChatGPT confuses a temporal regress with an explanatory regress. Remember we differentiated those in the kalam argument and the contingency argument. ChatGPT jumps too quickly to talking about a temporal regress of explanations. But what's really interesting here, I think, is that it doesn't seem to allow for unexplained brute facts. It either has an infinite regress of explanations or it terminates in a necessary being of some sort, but it will not allow that there can simply be unexplained brute facts. And that's hugely significant.

KEVIN HARRIS: Next, ChatGPT offers a possible way to avoid the infinite regress. Clip number six.

CHATGPT: Another perspective some philosophers consider is that the chain could be circular or looping, though that idea is less common and more controversial.

ALEX O'CONNOR: How could a series of contingent causes loop in a circle? Like, say my microphone existing if there's a contingent series of causes that explains why that microphone's there in the first place. Surely one of those causes can't be the existence of the microphone. It can't come full circle like that.

CHATGPT: You're right. For something like your microphone, it wouldn't make sense for its existence to be part of the cause of itself. That's why most discussions around contingent causal chains either lead to a necessary foundation or consider an infinite regress.

KEVIN HARRIS: Alex is not buying the eternal loop thing. That seems to be the main alternative, at least the one that the program brought up. Are there any others you can think of?

DR. CRAIG: The main alternative would be just to assert the existence of brute facts – that there are certain things that just are contingently without explanation, and ChatGPT is not allowing for that alternative.

KEVIN HARRIS: Here's the next clip. Alex makes ChatGPT stick to the facts. Clip number seven.

ALEX O'CONNOR: Is it then a fact that the explanation for the existence of this microphone has to either be an infinite series of contingent causes or some kind of series of causes that bottoms out in a necessary truth? Are those the only two options?

CHATGPT: Given what we've discussed and the logical constraints, those are indeed the two primary options typically considered.

ALEX O'CONNOR: OK, but remember I'm looking for facts here. I don't really care what's typically considered. I want to know: Is it true that those are my only two options?

CHATGPT: Got it. Focusing on the facts, based on logical analysis and the principles of causation, those two options, infinite regress or a necessary foundation, are the most coherent possibilities given our current understanding of facts and logic. While there are speculative or unconventional theories, they don't hold up to factual scrutiny in the same way. So, yes, those are effectively your only two solid options.

KEVIN HARRIS: Alex won't let the program speculate. This is kind of scary to me. He's being very personal with this machine to make it stay with the facts.

DR. CRAIG: Yes. Now, notice that the machine presupposes something like the principle of sufficient reason – that every contingent truth has an explanation of its truth. That is the common presupposition behind this dialogue so far.

KEVIN HARRIS: Well, we have a lot more to get to. And by the way, if you like free stuff and you like to sharpen your mind, then you're going to like Equip from Reasonable Faith. The Equip platform is our on-demand training center for Christian apologetics and philosophy. Now, whether you're new to the conversation or ready to tackle the deep stuff, Equip gives you the tools to defend your faith with confidence. We have courses on everything from Apologetics 101 to Molinism to the attributes of God. You name it, we got it. Just go to KnowWhyYouBelieve.org or you can go to ReasonableFaith.org to create your free account and start learning today. Let's pick it up right there next time on Reasonable Faith with Dr. William Lane Craig.^[1]