April 3, 2023

Sanje: Welcome everybody, I am delighted to have on the Divine Learning Podcast my friend, Jesus Christ, one of the big three in Heaven, commonly known as the Trinity, to discuss A.I. and its impact on learning. Is it OK to call you my friend, Jesus?

Jesus: Of course, I see all my followers as my friends. Friendship with me is at the core of what it means to be a Christian. Very glad to be here and the other “two” of me (God the Father and God the Holy Spirit) send their best; of course, they are here in spirit, if you know what I mean [winking].

Sanje: Haha. So let’s get at it. Before we delve into A.I. could you please give us a sense of your approach to learning in general, and how it works best?

Jesus: Well, obviously I have some thoughts on how best to learn. But let’s be clear. I personally don’t learn, or better put, I cannot learn. That is probably a little difficult to comprehend but think about it for a second.  If I did learn, I might wake up like you; check my phone, see a headline in the news, and find out someone else is now God. Like Yeezus? You might say that it makes me a “Know-it-All”, but I prefer omniscient [Jesus winks again].

Sanje: Fair. I finally get the whole predestination thing now. So I can respond to the “how do we know we are not just puppets?” free will objectors.

Jesus: I think my biggest point about learning in general, and I made this time and time again 2,000 years ago, is that learning is about knowledge creation, what you know rather than what you repeat or perform in the temple or church. There is a lot of performative stuff that looks like knowledge. To get to knowledge there must be both thinking and action and a catalyst, such as reason or faith. The transformative reaction that results is knowledge, and that is an emotional experience in itself. Knowledge without emotion is shallow, more like information.

Sanje: So how does all this relate to the recent human awe around Artificial Intelligence, ChatGPT, etc?

Jesus: There are quite a few things to consider with A.I. I think the first is somewhat obvious. When you are constructing knowledge you need to develop a mental (social and emotional) model of the person that is responding to you in the learning process. Just like when you receive guidance from any source in the construction process—a parent or a friend or a teacher—before they have opened their mouth you have already ascribed certain attributes to them. They are authentic or they lie a lot, they care, they never read, or they always take this side or another—these are their attributes in your mind and they create the foundation for trust, the key requirement for learning. One of the underlying assumptions of that trust is that what you say may affect how they respond. What are the equivalent attributes of your A.I. bot? That should be concern No.1.

Sanje: You are reminding me of my favorite quote by the famous Christian author of the first part of the 20th century, A.W. Tozer who said “man’s spiritual history will positively demonstrate that no religion has ever been greater than its idea of God”. He was making the point that Christians must be able to articulate what they thought were the attributes of God. What is His, well, yours, in this case, personality and character? It would show how they had experienced these Godly attributes. If they were stumped, it would say something about their growth and relationship with their God. The problem then for me with A.I. is that these bots potentially possess the aggregated attributes of everyone whose thoughts and knowledge they have consumed. But how does having a condensed version of conventional wisdom help advance knowledge, unless that’s all you want, I suppose? You will never find their attributes (some might call them human biases) by just looking at the algorithm as you will be consuming all of them based on their ingestion and digestion profiles. 

Jesus: Yes and one of my attributes is a bias toward Truth. And that brings me to my second point. A.I. only has one action and it’s one way—responding to your prompts. But are you responding to any prompts from it? 

You can tell my attributes from all my actions and responses, which leads to “knowing” me. Do I answer some prayers but not others? Do parents heed every request? Those are prompts. Similarly, you know how your teacher in your classroom acts and you also know whether you can trust them by the way they respond to your failures and your response to their feedback. If they never or rarely respond, they are expecting something of you that is more akin to performative learning rather than knowledge. If you don’t respond, you are probably not learning. You know my approach to failure, it’s an opportunity, and that is true of most learning. Knowledge creation has to go two ways. One person cannot be doing all the prompting. That’s poor pedagogy. Why do you think I responded to the Pharisees with questions of my own or parables? To force thinking and a response—and generally speaking they did not respond, except to recoil. If it’s only one way there is no real experience because there is no trust. Would you continue to trust someone who never responded with prompts of their own? That would be madness or superstition or empty worship or the wrong “God”, what I call idolatry. 

Didn’t I hear Alexa and Siri never managed to fulfill their dream of prompting you to order stuff but succeeded in getting you to just turn on the music and the lights? If the prompting process is only one way (you use their responses in a performative way), then it’s unlikely to lead to trust or learning and ultimately knowledge. So how consistent and reliable are the actions of these chatbots and what prompts did they give you?

Sanje: Wow. I now get the scariest thing you ever said in the Bible, which keeps me up at night sometimes. I think it was “But I never knew you”. Knowledge is a function of a two-way experience. If you don’t know me, I probably don’t know you. It’s more than lurking on Facebook or LinkedIn as a connection. I am acting on my own and I am trusting myself rather than any of the sources that I have been exposed to. I am basically deceiving myself that I have a relationship with you, and actually, my real trust is in some other place such as money, comfort, or pride. Got it. What else?

Jesus: I think to be honest the biggest issue, and my third point here about A.I., is one of context. With A.I. the ability to provide context is severely hampered by its inputs and its design. We just don’t know the context of its responses. And to be fair this is what has led to some error, confusion, and division in Christianity itself, too. When I say context, I also mean the unity of the conclusion, not just its weight. 

Let me give you an example. You remember the Apostle John is quoted in the Bible as saying “God is Love”. Now obviously We (the Trinity) are Love. But that does not mean We are only Love. We are also a whole host of other things: We are truth, We are holiness, We are grace, We are omniscience, We are self-sufficiency and many more attributes. One cannot cancel or minimize any of those attributes because it would make us less God, and in some way not self-sufficient. Does your God need something? Does He need to be, say on this issue, less Truth per what He has explicitly said and more Love? That is problematic. Frankly, we don’t need anything, not even humans, to accomplish our will. We also require (not need) a whole host of other things to develop a two-way experience, such as reliance, obedience, and more. So if you thought just reading that bit about Love and basing your whole world view of what We, as Christians, stand for, you might be disappointed; it’s not just loving one another, nor is it just doing loving acts (“you are saved by faith through grace, not deeds”), etc. It is knowing all our attributes. If you also look at the theological and political conflicts that have broken out about Christianity, it is all about one area of over-emphasis rather than another justifying some practice: who should lead prayer and service, who should be allowed to have a relationship with me, who should get married, etc. 

How do you then arrive at that right context, the conclusion that represents good context? Well basically, as I said earlier, the organs of faith and trust through which your experience is real. You will be prompted to work out what is right. Sometimes those prompts will be gentle like nudges, sometimes a gentle push, and sometimes so loud they are inescapable. From experience, not intelligence. “I really did know you” becomes the goal and the solution. 

Sanje: Yes it comes down to reliability or trust, and in fact, one of our team recently wrote about this by asking the bot itself whether it was reliable. I think it is a little difficult for us humans to think of knowledge creation as a continuous bilateral experiential process. I would venture that it is because humans have strived to make learning bite-sized, simplified, and standardized: less collaborative, less social or emotional, and very performative. A nicely packaged sermon with insert-the-blanks, a verse of the day, etc. This is all very thought-provoking. 

So one final question, a light-hearted one to finish. I can see the development of a little culture war about A.I. on the horizon over which gender these bots should be: many have neutral names like Sydney, and others have female names like Alexa and Siri. Any thoughts on this? 

Jesus: I think you are trying to trap me, Sanje. Just like you know who a few thousand years back.

Sanje: Fair. Sorry. Keep it tongue in cheek then, you have a great sense of humor. After all, you created humans. But please, no parables.

Jesus: Hmm. I would think as regards gender that they should be…not male, not female…binary perhaps?

Sanje: And that was even better than a parable. That’s a wrap for this week on the Divine Learning Podcast. Thanks to our guest today, Jesus Christ. He is literally everywhere these days! We hope to see others from above in future episodes, giving their take on current issues in education and learning.