Pope Leo XIV’s recent encyclical on artificial intelligence, Magnifica Humanitas (“Magnificent Humanity”), calls on political leaders to develop policies that address the most dramatic effects of AI on human life. But at its heart, it is a deeper reflection on what it means to be human at a time when artificial systems can imitate human intelligence with increasing realism. As AI creates an illusion of the “fake human” it raises a critical question: What truly distinguishes a human being from everything else? What makes you different from an inanimate object like the seat you are sitting on?
You are made mostly of water and a very special cocktail of other atoms, but the seat is also a special atomic cocktail. If it is made of wood and natural materials, its composition is surprisingly close to yours, though maybe not so surprising since we are what we eat — and we eat plants. But you are reading this and the seat isn’t. You have a natural intelligence, something that the Apple cofounder Steve Wozniak called “actual intelligence” at a commencement speech in May. And that, to a large degree, is what makes you human.
In the flurry of our lives – work, meals, family – it is easy to lose wonder at this wondrous thing and to take it for granted: that intelligence emerged from your particular atomic arrangement just as it didn’t for so many other arrangements. But it did, and it is yours to do with as you wish.
The emergence of AI has profoundly shifted things, prompting the Papal encyclical, because AI is such a good mimic of our natural intelligence that we are figuratively, and sometimes literally, seduced by it. Here is a machine that I can talk to as if it were a human being – the literal definition of the Turing Test for intelligence developed in 1950 by the mathematician and computer pioneer, Alan Turing.
The latest AI is an amazing development and an amazing opportunity, but is it also a threat to our very humanness? AI can take away tedious thinking tasks from us such as scrolling the internet, writing emails or software code, and help us increase productivity or free up time for recreation. But it can also lead to job losses and even the destruction or enslavement of mankind. And based on all the boos for this year’s commencement speakers who extolled AI, the kids are not alright with it. If things feel a bit out of control, it is because they are, and our leaders need to take note. But as individuals, we can empower ourselves, and renew our relationship with the human within us, in part by actually learning from how AI learns.
Large language models were designed to mimic actual brains made with neurons and synapses that connect them. The learning process involves updating the strength of the connections within the complex network of artificial neurons. In the artificial brain these connection strengths are just numbers stored in an array. The LLM learns the best numbers such that when input prompts are fed into the network and propagate through all the weighted connections, meaning emerges.
But how is it done in our actual brains? Scientists are just now at the beginning of an understanding of brain function, but if LLMs are good models for our brains, and good mimics of our intelligence, we can understand that the neuronal connectivity, and the strength of those connections, are important. In our human brains, these are not numbers stored in an array. They are real physical connections, made of atoms. When we learn, to update our network connections we have to move atoms around and arrange them in very particular ways. This takes energy and effort. This is why, when we learn, we get hungry and tired, and why learning requires deep concentration and struggle. This cannot be bypassed. Every time we get lazy and short-circuit this process by asking AI to tell us the thing we want to learn, we miss an opportunity to rearrange atoms in our mind and thereby to learn. The struggle itself is part of the learning, taking great awareness and concentration. Our own determination to fully engage with and sense the world and interpret the results is in large part what makes us human.
We still have control over how we use AI and interact with it. Using AI to take over tedious thinking tasks is fine if it frees up time to learn something new that we are passionate about, or indeed if it helps guide us in our learning struggle. And relaxing with your phone can be a reward after a hard day. But when we use AI to short-cut our actual learning, or use social media content to replace our sensed universe, then we are letting the tech make us a bit less human and a bit less magnificent and more like that seat we are sitting on.