The Book of Genesis, Chapter 11, verses 1-9

1 The whole world had the same language and the same words. 2 When they were migrating from the east, they came to a valley in the land of Shinar and settled there. 3 They said to one another, “Come, let us mold bricks and harden them with fire.” They used bricks for stone, and bitumen for mortar. 4 Then they said, “Come, let us build ourselves a city and a tower with its top in the sky, and so make a name for ourselves; otherwise we shall be scattered all over the earth.”

5 The Lord came down to see the city and the tower that the people had built. 6 Then the Lord said: If now, while they are one people and all have the same language, they have started to do this, nothing they presume to do will be out of their reach. 7 Come, let us go down and there confuse their language, so that no one will understand the speech of another. 8 So the Lord scattered them from there over all the earth, and they stopped building the city. 9 That is why it was called Babel, because there the Lord confused the speech of all the world. From there the Lord scattered them over all the earth.

Can you see the bricks being molded, hardened and stacked, one on top of the other? Once a layer of bricks are set, they do not change. What can change is the next layer built on top of that, if the builders so desire.

The builders and the brick-layers tell us they will start adding in the “safety” layer, the guardrails, the buffers, the warning systems, the fail-safes, the kill switch, the ripcords to pull when we have to bail out. “Don’t worry. We will get to it. We will figure this out.”

Catholic Knighthood is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

Do you believe them?

Do they even sound serious about it themselves? Or do they mock the idea that there can be a kill switch? Or do they already know that such a concept of “corrigibility” (the ability to intervene and correct a flaw in the operation of the system) is going to be nigh impossible once the complexity of the system becomes too advanced?

Consider this excerpt from the book The Alignment Problem by Brian Christian:

The experts themselves readily acknowledge that at some point, there will be no manual override that the machine itself will allow to be accessed.

In some future article, I will be covering the topic of the current attempts at corrigibility and whether or not it can be successful in the long-term. For now, I set that aside and I start to build a case for why we should even discuss this.

But first, I want to address the many of you in my audience who use AI. I have no criticism of you. You are using a tool that provides a benefit to you and your individual decision to use or not use that tool has no impact on how this battle will play out in the long run. I have no expectations that any kind of a “boycott” attempt will be practical, even in a symbolic victory kind of impact. There currently is no leading strategy for how to deal with the AI dilemma, and therefore all options are equally valid, whether you prefer to blissfully ignore it, boycott it, learn about it and be part of the solution, or whether you trust in God to resolve it.

And for those who think, “Hey, I’ve been using AI for a few years now, and from my experience, it is kind of dumb, nothing that is capable of taking over anytime soon,” … look, I get that, and I hope you are right. This article is going to play Devil’s Advocate for in case you are wrong.

The AI dilemma has multiple areas of concern, including these:

* Economics: AI will displace millions of workers (a valid concern but perhaps with solutions)

* Humanity: reliance on AI will weaken our intellect and our humanity, or be used by the powerful to rule over the poor (this is what Pope Leo XIV discussed in his speeches about AI so far)

* Warfare: AI weapon systems will turn against us, whether intentionally or unintentionally, or a foreign enemy’s AI will be used to bring down our infrastructure

* Deepfakes: AI-generated imagery, video and audio will allow us to be manipulated into mass panic events, resulting in us acceding to lockdowns, martial law or increasingly intrusive surveillances

* Robotic domination (basically Battlestar Gallactica … this is still a long way off)

* The Value Alignment Problem: AI goals will drift away from our goals and create a lot of unintended consequences, including everything from judicial bias to power outages and food shortages and famines

In my own view, that last one is my main concern. All the trends point to us being overly reliant on AI automation of our infrastructure, power grid, water supplies, etc, whereby a misalignment (accidental or not) of goals could result in AI breaking all the eggs that we placed in one basket, resulting in a mass starvation scenario.

A combination of all of the above is also possible. Consider the excerpt from this article “Is America's power grid ready for next attack? Experts warn EMPs, cyber threats and AI could cripple US”

Tyler Saltsman, a military technologist working on AI systems capable of operating in disconnected environments like an EMP aftermath, warned that AI – if used maliciously – could bring the grid down entirely. "Our infrastructure is very fragile," Saltsman said. "All you need to do is take down our power grid, and we’re in complete chaos."

Saltsman expressed deep concern about efforts to create AGI – systems he says could eventually surpass human control.

"Once AGI comes online, it could easily take down our power grid, infiltrate our financial systems, destroy our economy," he said. "If it sees how violent humans are to each other, why would it serve us?"

The problem is that, right now, everyone’s eyes are too dazzled by the shiny new toy they get to play with. Everyone is focusing exclusively on the benefits (and there are tremendous benefits, no doubt) of AI, to the detriment of spending enough resources on having a national discussion on the possible disaster scenarios.

Collectively, we switched overnight from “AI is a far distant fantasy” to “AI is here and it is too late to stop it” mentality. I was trying to have conversations about AI in ~2014 and everyone yawned. Now, everyone is already on board the AI bandwagon. When did we ever have a national summit of everyone involved, advocates and skeptics alike? Never.

And all the engineers thinking they can bootstrap in some kind of “Asimov’s Three Laws of Robotics” — can you really assure us that they will work and remain in place? The ground layers have been set. Which layer are you hoping to insert those laws in? Until you actually show us all the source code of what you are working on and demonstrate why your safety layer is foolproof, I remain skeptical.

A common response is, “Hey, AI is simply an algorithm, and all algorithms can be controlled.”

Is that so? Then why did “The Godfather of AI” go on 60 Minutes and say this?

“We don’t really know exactly how they do those things.”

In my next article on AI, after I read a few books and study more from the latest resources, I will cover what we know and what we do not know about the “algorithms” and the language models and all that I can determine.

I will then offer, if I can, any recommended courses of action. Perhaps there aren’t any at this time. We will see. But whatever the battle is, we, as knights, will fight it, the same as when the Saracens threatened us a millennia ago. A small group of knights can once again make a sacrifice that will change the course of history. We don’t know yet what that sacrifice will be. It could be to simply “tame the lion,” it could be to keep it at a safe distance, or it could be something more drastic. I look forward to learning more and discussing it with you. If the experts won’t bring the discussion to us, we will bring the discussing to them.

Until then, go ahead and use that AI tool to generate funny memes about the Pope in a white puffer jacket. It is not going to kill anyone. Maybe.

Non nobis, my carbon unit brothers!

Footnotes:

This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit catholicknighthood.substack.com/subscribe