Almost Intelligent?

ACT member Huw Rees shares some of his own reflections on navigating artificial intelligence wisely as a Christian. In this piece, he discusses his experiences using AI while developing his educational series Science Shorts. You can explore these valuable free resources at the Science Shorts YouTube channel.

We’re always delighted to showcase the work of our members—if you have something you’d like to contribute for consideration as an article, please do get in touch.

“Here’s the list you asked for – you can copy it, or click this link to download it!”
“Grok…did you make up that link?”
“You got me – I did make it up! That was an example of what I could have done if this was a real-life scenario.”

Yes, Elon’s AI that he claims is going to reach AGI status (General Artificial Intelligence – smarter than a human) anytime now still doesn’t know the difference between real life and simulation. That raises a deeper question: can circuitry ever truly ‘know’ anything? Granted, the list it made for me was exactly what I needed, but all it was doing was using clever maths to calculate the next most probable output needed. Genius, and yet still unthinking.

When I left teaching two years ago (see my previous article “The ‘War’ On Truth”), I was fortunate enough to have a couple of other ventures to fall back on – one of them being my educational YouTube channel, Science Shorts. My brain was never very good at retaining a lot of information, so my forte has always been the ability to distil the huge amount of information students (and I) need for the Sciences into a more manageable and memorable format. I tried to get various AIs to help me with writing scripts for videos, but it just never worked out – they can’t know what bits to emphasise, what bits to spend less time on as they rarely come up in exams, for example. At least, not as well as someone who has years of experience. While they were useful in checking what I had written, I gave up on them having any worthwhile input before I was done. (I should say that no AI was used to help me write this!).

This backed up my long-held assertion about AI in general: it’s incredible at spotting patterns that we can’t see (for example, being able to detect breast cancer in mammograms with incredible accuracy when even doctors don’t know what it’s picking up on), but it can never truly create anything – not in the truest sense of the word. Rather, it just takes an aggregate of all the information it has access to, and churns out an average.

Coca-Cola (justifiably) recently came under fire for releasing their yearly Christmas (or ‘holidays’) advert…made completely with AI. It’s almost immediately noticeable when you see the uncanny valley, glassy-eyed cartoon animals that will likely make you feel intensely uncomfortable for a reason you can’t put your finger on; but it’s when you realise that the dimensions of the Coke trucks sliding through the winter wonderland are changing shape every few seconds (even missing all of their rear wheels at points), that you begin to question the sanity and observational skills of the people who greenlit the ad (oh the irony of “it’s always the real thing” being sung in the background!).

Similarly, a Christmas mural was very recently hung in Kingston-Upon-Thames, depicting a festive scene of yesteryear. The issue? Look (not even that) closely and you’ll see melted faces, several extra fingers on hands, and dogs so contorted that they wouldn’t look out of place in John Carpenter’s The Thing. It was torn down shortly after being rumbled as AI ‘slop’.

Hilarious as those examples are, what happens when the consequence of AI making the wrong decision is literally a matter of life and death? A notorious example that’s become a meme in its own right is the fact that an AI will tell you that a mushroom is edible if you show it a picture of it, only to forget to mention that it is, indeed, poisonous (Skynet would be proud). Google’s Gemini went through a phase of telling users to self-harm after asking it perfectly innocuous questions. A fellow in my church is involved in developing the decision-making processes of self-driving cars; the problem being that they actively have to program aggression into it – otherwise it could end up sitting at a fairly busy junction for half an hour. But who’s responsible when it takes it too far and causes an accident? This is one of the reasons why, as I’ve long held, AI will have (and already has had) a huge impact on our lives, but it will never supplant our God-designed minds; and that’s before taking into consideration our spiritual nature.

So, you can see why I steered clear of using AI for a long time. However, my viewers started asking for a more interactive way of testing their knowledge, so I approached a talented programmer, David Smith of Made Responsively. He recently released a Christian Dream Journal app, which uses AI to help believers keep track of their dreams, with the aim of encouraging and giving suggestions for points of prayer; but never to interpret, as that’s the role of the Holy Spirit. He subsequently did a fantastic job creating an app to help students revise with the quizzes I wrote; even implementing an AI feedback feature, which gives them specific pointers on how to improve. It’s freakishly good, but not perfect, as it can still have the tendency to argue with itself from time to time!

Having had many conversations on this topic, it seems a huge number of people are excited about the possibility of AI becoming smarter than humans – I’ve been trying to figure out why. I think that for many it’s a case of wanting proof that the most mysterious aspect to our biology – our brains – are merely good computers; that there is no divine spark within us; that we are not created, but just entropic matter floating through space (that came from nothing…somehow). “If ChatGPT simulates my thought processes perfectly, then there is no God, which means I get to live my life the way I want.” That might seem reductive, but I really do believe that this is the way many people think, if not subconsciously. I think it’s a similar story when such people get excited about the possibility of life on other planets: “See? We’re not special after all!” (It still wouldn’t prove the absence of God, of course!).

“I think, therefore I am” (Descartes) is in my opinion one of the most philosophically profound statements ever uttered; but the emergence of AI, that gives the appearance of thinking, prompts an even deeper question: what even isthinking?! I know I think – I can reason with myself; I have an inner monologue; I can rotate an apple in my head (look it up!). But how do I know others do too? Some people say they can’t do all of these things – what does that mean? Do animals think? Do single-celled organisms think? (After all, they behave like they do) or are they just chemical robots? (Is there even a difference?) And if they are the latter…could some people be too? It’s been proven categorically that the prompts being used to direct the biggest AIs are created with social engineering in mind: to promote certain (Godless) ideologies, and suppress others (but that’s nothing new for the internet). But if I can’t trust it; how do I know I can trust my own thoughts?

Great – all I wanted was help with writing a Best Man’s speech, and now I’m up all night questioning reality.

You might be tempted to draw the parallel between God making our minds, and us making the ‘mind’ of AI – but it’s just not the same. AI is essentially just electrons jumping up and down in semiconductors, while we’re told in Genesis that God needed to breathe the breath of life into Adam, despite all the required physical components being present. We are more than the material; don’t be drawn in by the (to some) appealing lie that we are purely physical beings. The Lord says in Isaiah 55 “…my thoughts are not your thoughts, neither are you ways my ways”. We should rightly be humbled by this fact, but we can also be encouraged by the fact that we think ‘as God does’, albeit on an infinitely smaller scale, and with sin infusing every ‘prompt’. We are made in His image, after all.

Can you therefore trust your own thoughts? Absolutely not! Don’t try to rationalise an existence that is, thanks to sin, incredibly irrational a lot of the time. Don’t look for the answers to life’s big questions in AI, or any other creation; look to the Creator who operates beyond the material for what you need. Don’t be concerned if it seems that others, including AI, ‘think’ that your faith has no basis (ChatGPT seems to believe in God though, funnily enough!) – their thinking is flawed; just as yours is. “Ah but that’s circular reasoning!”. Well, all reasoning is circular in the end. The only thing you have to decide is where you put your faith: in your own understanding, or in One who gave it to you in the first place? Ultimately, that’s all that our Christian walk boils down to.

Huw Rees