In a previous blog post I explained how the development of artificial intelligence has not reached the level of systems like HAL in 2001: A Space Odyssey. In the 1968 movie, HAL is an the artificial intelligence that controls the systems onboard an exploratory space craft. During the mission, HAL makes a mistake and the human crew decide to disconnect it. This leads to all kinds of disaster as the AI decides to fight back to preserve itself and the mission.
These are the kinds of scenarios most people fear when they think about artificial intelligence. Even though we aren’t anywhere near this level of AI yet, the possibility is troubling. “What if we lose control? What if it turns on us? Will we be able to stop it?” These and similar questions have been discussed and debated on YouTube, Facebook, and other social media platforms. I won’t rehash those discussions here. Instead, I would like to reflect on some of the dangers our current AI capabilities hold.
There is the danger of laziness
AI will tempt some to laziness and ‘cheating’. Plagiarism is not a new problem, but with AI it can be taken to new heights. With the help of AI like ChatGPT students, scholars, or pastors can write thousands of pages of content with minimal effort. I preached on AI recently and asked ChatGPT to write the introduction to the message to make this very point. It didn’t sound like me, but it wasn’t half bad. In case you were wondering, it fooled a few people, but I owned up to it immediately.
In a way, using AI to write for us is an advanced version of “copying and pasting” from your favourite website. Only now, with the help of AI, the material can be reordered so that it will appear original. Some have even compared it to ghost writing: employing another person to write on your behalf, and then taking the credit for the product. Only in this case, you won’t have to pay the ghost writer! Either way, the one taking the credit didn’t do the work.
The Bible has quite a lot to say about the sin of laziness. Proverbs warns the sluggard that “poverty will come upon you like a robber” (Prov. 6:9-11). Instead, we are called to be good stewards of the gifts, talents, and time that God has given us. We are to do “honest work with [our] own hands” (Eph. 4:28), such that would honour God (Eph. 6:6-7). Remember, AI is a tool that can assist us in our work, but it should not be used to replace our work. If AI assists us in doing our work more efficiently, this means that we’ve been given the opportunity to do more, not less.
There is the danger of fake church
David de Bruyn recently highlighted this danger here. I raised a similar concern during a recent message on the topic of AI and the church. We already have the problem of “market friendly” churches that focus on appealing to the widest audience, often at the expense of biblical truth. The rise of AI might add to this phenomenon and do away with the biblical preacher too.
Artificial intelligence can already create fake voices and faces, sometimes called “deep fakes.” These digital likenesses of real or imagined people can be disturbingly convincing. Combine this with ChatGPT’s ability to write sermons and you have the potential to create the perfect digital preacher. Just choose the voice you like listening to the most (James Earl Jones perhaps?), pair it with the face you find most appealing, and have it read a sermon written in the style of your favourite preacher… voila!
You may think that this is an overreaction, but this is exactly what a German church scholar did last week. You can read about it here. The artificial preacher, which presented as a Black man with a beard above the altar of St. Paul's Church in Fürth, Bavaria, told the packed congregation not to fear death, according to the Associated Press. The service wasn’t flawless and some congregants reported being put off by the artificial manner and tone of the “preacher,” but it is only a matter of time before these hurdles are overcome.
Many Christians are opting for online church and preaching, and this has only been accelerated by the pandemic. Worship is viewed as a passive event and even those who attend in-person services often leave before anyone else has a chance to greet or engage with them. We’ve fallen into the trap of treating our worship like a form of entertainment.
This is not what God has called believers to. The qualifications for elders describe godly men (1 Tim. 3:1-7; Titus 1:5-9), not robots. They will be imperfect, but they have been given the responsibility of “keeping watch over your souls” (Heb. 13:17). There is more to being a pastor than preaching. The name implies shepherding, which is impossible to do without knowing the sheep.
AI preaching would not only isolate believers from biblical oversight, but also from one another. God’s design for the church is community, which means that we have fellowship with God and with one another (1 John 1:7). There are numerous “one another” commands in the New Testament, all of which would be impossible to fulfil in isolation. The perfect church would not mean the absence of people, but would be filled with imperfect people loving and serving one another to the glory of God.
These are two very real temptations that we have to contend with in our new AI integrated world. A biblical view of work and of church will help guard our hearts against them.
Artificial intelligence is trending. News articles, documentaries, and discussions on the topic abound. What was once considered science fiction has become mainstream. Even those who’ve never heard of Isaac Asimov are suddenly interested in what shape our AI-integrated future might look like. One reason for the surge in interest is undoubtably the rise of ChatGPT.
ChatGPT or Chat Generative Pre-trained Transformer is an artificial intelligence chatbot. We’ve all encountered chatbots, whether it is a “How can I help you?” pop-up on a website or an automated reply service. What makes ChatGPT unique, however, is the natural language processing it uses to create humanlike dialogue on almost any topic. It was first released in November 2022, and has only improved since then.
Many have touted the arrival of ChatGPT as another step towards a technological singularity: a hypothetical moment when technology advances so far that it will become uncontrollable and irreversible. This is the stuff of science fiction nightmares. Even though we have come a long way, we aren’t anywhere near the dystopian disaster many fear.
Artificial intelligence can be categorized into three levels. The first, limited AI, includes all artificial intelligence programs or protocols that perform a specific task. It may do that task well, but it can’t do anything else. Deep Blue, which beat world chess champion Garry Kasparov in 1997, is an example of such an AI. It is very good at chess, but not much else.
The second level, general AI, would be an artificial intelligence capable of a wide range of functions. It would be able to learn new functions and improve existing ones. It would be almost human, or at the very least, to most observers it would seem human. The third and final level, super AI, would be an artificial intelligence capable of improving itself beyond our human capabilities, surpassing our intellect.
While some believe that we are only a small step away from general or possibly even super AI, the reality is very different. Currently, our most advanced AI (those that employ machine or deep learning) are nothing more than capable but very limited tools. They aren’t anywhere near the level of sophistication that would make them “almost human.”
How should Christians reflect on the advance of artificial intelligence? Perhaps it would be wise to start with the humans who create artificial intelligence. There is no way to remove the human element entirely. A human planned and programmed the processes that the AI uses. A human created the data that advanced AIs like ChatGPT are trained on. A human gives the AI the instructions that it executes so well.
This has two immediate implications. First, when we understand that humanity is central to the development of artificial intelligence, we also understand that sin is close at hand. An AI, like ChatGPT, has no way of knowing whether the information it was fed, is true. It cannot account for the biases of those who created it or the data it was trained on. It can write the lyrics to a Christian worship song or imitate your favourite author, but it does not know whether what it generates is good or true. That depends on the data, the user, and the algorithms, i.e., humans. Trusting in AI or hoping that it will be the miracle that saves the world, would as foolish as trusting in man.
The second implication is related to the first: it should make us appreciate the unique glory that God has bestowed upon humanity. Even though we are fallen creatures, we are still said to have been created in God’s image (Gen. 1:26-27). Each attempt to create artificial intelligence, to create something “in our own image,” reminds us that we are unique. The closer we think we get to an artificial intelligence that matches our own, the more we come to appreciate the difference between us and our machines. We have a capacity for the good, the right, the beautiful. We are more than sophisticated robots, a collection of chemical reactions and firing electrons. We are image-bearers of the most High.
Like with most advances in human technology, there will be triumphs, fears, and failures. For now, at least, we can rest easy that we have not created something “human”; only God can do that.