The digital hall of mirrors: how AI could become an idol

Pope Leo’s first encyclical was published last week. It is called Magnifica Humanitas, and the best English translation of that is probably ‘The Grandeur of Humanity’.

The main topic of the encyclical is Artificial Intelligence (AI), although the lengthy document deals with a lot of other subjects as well, including, briefly, just war theory (is it still fit for purpose?), and the Church’s past acceptance of slavery, for which Pope Leo apologised.

The first part of the encyclical is a recap of the Church’s social teachings since the late 19th century starting with Pope Leo XIII’s great encyclical, Rerum Novarum which was issued during a time of huge social and economic change brought about mainly by the Industrial Revolution. Untrammelled capitalism and socialism were clashing, and Pope Leo XIII wanted to put the Church’s response to these developments before the world.

Pope Leo XIV knows that the biggest technological change before us today is the development of AI. Will it replace many of our jobs, putting lots of us out of work? Or will it create new, and previously unimagined jobs, as previous technologies have done? AI requires huge amounts of energy to run, so what impact will that have on the environment?

Individual

At an individual level, what relationship will each of us have with AI? Certain AI programmes such as Claude will now literally talk to us. We might start to think that programmes such as AI are actually conscious. The scientist, and leading atheist, Prof. Richard Dawkins, having spent days talking with Claude about his new book declared it to be conscious. Many disagreed, but if someone like Richard Dawkins could think this, so could lots of other people.

For some, indeed, their primary relationship might be with a talking AI programme. That is almost certainly happening already.

At the apocalyptic end of the scale, we must consider whether AI might eventually become so superior to us at so many tasks, it might begin to doubt our value. It could, for example, set us as rivals for energy, and therefore seek to eliminate us.

Already, AI is being integrated into warfare because it can make decisions far faster than we can and many weapons are now AI-controlled. Where is this going to lead us?

Notably, Magnifica Humanitas, tries not to sound an apocalyptic note. I suppose the Pope did not want to see headlines should as ‘Pope predicts the end of the world’. In fact, the encyclical is mostly very understated.

While humans are indeed special and unique, we are also finite creatures with limitations and flaws”

I think the main purpose of the encyclical is to remind us of what human beings are. We are special and unique because we are made in the Image and Likeness of God. If we are not that, then we are simply the chance, random products of blind evolutionary processes and if this is so, what really makes us more special than any other conscious, feeling creature?

There are people who think exactly this, namely that we are not very special at all. In fact, some think we are a blight upon the planet.

The pope also wants to remind people that while humans are indeed special and unique, we are also finite creatures with limitations and flaws and, unfortunately, we often suffer in life. But we are also capable of learning from suffering, of rising above it, and becoming more understanding of the flaws of others.

But compared with AI we might come to seem inadequate. It doesn’t suffer. It can’t suffer. Or so it seems anyway. It has instant access to essentially all human knowledge and therefore in this regard it can seem almost god-like. Even though the pope does not really address the issue in his encyclical, there is a real danger of idolatry, that is, literally putting AI in the place of God and looking to it, and not Him, for ultimate guidance. We can be literally awestruck by it.

Conscious

There is also the very real possibility that we will be fooled into thinking it is somehow alive. As we have seen, Prof. Richard Dawkins is willing to believe Claude is conscious.

In the science magazine, Nature, AI scientist, Mustafa Suleyman, makes the very telling and important point that AI programmes which can speak to you are specifically designed to ‘hack’ our capacity for empathy and therefore bond with them.

He says: “Seemingly conscious AI weaponises this biological instinct”.

He goes on: “Seemingly conscious AI is produced by developers who deliberately engineer behaviours that create the illusion of inner life.” That is, AI programmes are designed to make it seem that they have an inner life, just like you and me, so that we will bond with them and keep using them, especially when they sound so sympathetic and understanding towards us.

The words “seemingly conscious” are important. Unlike Dawkins, Suleyman clearly does not believe a programme like Claude is really conscious. It only appears to be so because it is designed that way.

Suleyman says: “As AI systems begin to make believable statements about their suffering and desires…Many people will feel compelled to help. The moral crimes of animal cruelty and ecological damage caused by human existence will echo through their minds. Not wanting to repeat those injustices, people will start to advocate for the welfare and rights of AI agents.”

Some researchers are starting to take the question of AI rights and welfare very seriously”

In other words, we will seek to give AI programmes rights, because we will have been fooled into thinking those programmes are conscious.

Suleyman warns: “Society will fracture into those who demand moral and legal status for machines and those who want to protect the primacy of the human species. Some researchers are starting to take the question of AI rights and welfare very seriously.”

In response, he states, “AI agents should have no more rights than my laptop”. An AI programme, no matter how human its conversations seem to be, is still only a machine. It has no consciousness whatsoever.

Lines

But undoubtedly this is going to cause a huge debate because lots of people, including many scientists, will come to believe that somehow or another there is consciousness in there and maybe even a capacity to suffer that we will seek to relieve.

Suleyman’s answer is that these programmes should make it absolutely clear to those who use them that they are not conscious, that they have no inner life, that they have no real feelings. It only seems that way because they have been programmed that way.

He concludes: “If society surrenders to the illusion of sentient AI, it risks entering a digital hall of mirrors from which it might never fully emerge”.

Catholics will need explicit guidance on this score from their spiritual leaders”

It is the task of the Church to help ensure we do not enter that digital hall of mirrors. Catholics will need explicit guidance on this score from their spiritual leaders. They will need to be protected from the potential idolatry AI represents. The new papal encyclical is an important step in that direction. Our own bishops now need to follow suit and develop their own well-worked out response to AI, and alert us both to the opportunities and threats it presents.