Artificial Intelligence or Just Mathematical Probability?

By Liam Lally | 9th June 26

I came across a Tweet (still call them tweets!) from Pope Leo XIV in which he said the following:

Artificial intelligences do not undergo experiences, do not possess a body, do not feel joy or pain, do not mature through relationships, and do not know from within what love, work, friendship or responsibility mean. Nor do they have a moral conscience, since they do not judge good and evil, grasp the ultimate meaning of situations, or bear responsibility for consequences. They may imitate or even simulate, but they do not understand what they produce, for they lack the affective, relational, and spiritual perspective through which human beings grow in wisdom. #MagnificaHumanitas

So I thought I would take such a profound statement and ASK an A.I. (Gemini) for a response…

The Pope’s statement outlines a view shared by the vast majority of computer scientists, philosophers of mind, and cognitive scientists today. From a strictly technical and scientific standpoint, yes, his description of current AI architecture is accurate.

The statement draws a sharp line between simulation (what AI does) and sentience/understanding (what living, conscious beings do).

Here is a breakdown of why this aligns with our current understanding of artificial intelligence:

1. Processing vs. Experiencing

AI models operate through complex mathematics, statistical correlations, and pattern recognition. When an AI generates a poem about grief or gives advice on a broken heart, it isn’t drawing from an inner well of sadness or empathy. It is calculating which words are most statistically likely to follow one another based on vast amounts of human-written text.

The AI: Lacks a physical body, biological nervous system, and subjective awareness (qualia). It doesn’t “feel” the data it processes.
The Human: Experiences the world through a embodied, biological reality where emotions, pain, and joy are tied to survival, hormones, and consciousness.

2. Mimicry vs. Meaning (The Chinese Room)

The Pope notes that AI can “imitate or even simulate, but they do not understand what they produce.” This closely mirrors a famous philosophical argument called the Chinese Room Thought Experiment (proposed by John Searle). (Read the whole paper here – https://rintintin.colorado.edu/~vancecd/phil201/Searle.pdf)

Imagine a person inside a room who doesn’t speak Chinese but has a massive rulebook that tells them exactly how to respond to Chinese characters slipped under the door. To the person outside, it looks like the person inside speaks Chinese fluently. In reality, the person inside is just manipulating symbols without understanding a single word. Current AI is essentially a highly advanced version of that room.

3. Morality and Responsibility

An AI cannot possess a moral conscience or bear responsibility because it lacks agency.

AI does not choose its actions based on personal principles; it follows code and optimization functions.
If a self-driving car causes an accident, or an AI system outputs biased data, the legal and moral responsibility lies with the human creators, deployers, or users—never the machine itself. A machine cannot be punished, nor can it feel guilt or a sense of duty.

The Philosophical Nuance

The only area where thinkers sometimes disagree isn’t about what AI is right now, but what it might become.

A minority of futurists and philosophers argue that if an artificial system ever achieves a complex enough architecture (perhaps through advanced neuromorphic computing or robotics), some form of machine consciousness or functional understanding might emerge. However, this remains firmly in the realm of science fiction and theoretical debate.

For the AI systems that exist today, the Pope’s words capture the reality: they are brilliant mirrors of human expression, but they lack the internal life, relationships, and spiritual weight that define human wisdom.

Within the responses of the tweet by Pope Leo XIV I found this comment an excellent, simplified description of what happens when you ask A.I. “anything”…

Exactly Right.

Regular people don’t understand and my IT professional colleagues unfortunately use anthropomorphic terminology that confuses them even more.

Any piece of software, including LLM AI technology, does not UNDERSTAND a single word.

Humans use words to denote a concept, but to AI, words are just bytes in a database.

consider the equation: Y = mX +b

For any given X, we compute the Y.

LLM is just like this, except the number of variables is extremely large.

However, all it does is compute Y based on the input.

The equation doesn’t understand what X actually is. It has no intelligence or conceptual ability.

All of this has been couched in anthropomorphic terminology either for evil purposes or simply to pump up the associated stocks.

So, back I went to my mathematical predictive text generator (sorry A.I.) and asked for a response…

Your colleague hits the nail on the head regarding the underlying math, and they are pointing out a massive, legitimate frustration within the tech industry: the hype machine and the danger of anthropomorphism.

Their breakdown of how Large Language Models (LLMs) actually work is highly accurate, though their conclusion about why we use these terms has a bit of cynical edge (which, to be fair, is often justified in the tech world).

Here is a look at why their response is so accurate, along with a tiny bit of nuance on the “why” behind the terminology.

The Linear Regression Analogy: Exactly Right

Comparing an LLM to Y = mX + b is an excellent way to demystify the technology.

In high school algebra, you have one input (X), one slope (m), and one intercept (b) to get your output (Y).

In an advanced AI model, you have billions or trillions of inputs and “weights” (the m and b equivalents), but the core mechanism is the same: it is a massive statistical calculation.

The equation Y = mX + b doesn’t “know” what numbers are, nor does it care if X represents dollars, temperature, or words. It just computes. Your colleague is 100% correct: to an LLM, the word “love” isn’t a feeling; it’s just a token represented by a specific vector (a string of numbers) placed in a high-dimensional mathematical space.

The Problem with Anthropomorphic Terminology

The IT professional is entirely justified in being frustrated with the language used today. We use words that imply human consciousness to describe mechanical processes:

“Learning” is actually statistical gradient descent (adjusting numbers until the errors decrease).

“Hallucinating” is just the model generating high-probability text that happens to be factually incorrect based on its training data.

“Understanding” is just pattern matching.

When the public hears “the AI thinks or wants or knows,” they assume there is a “someone” inside the machine. There isn’t. It’s just code executing on silicon.

Is it Evil, Stock Pumping, or Just Lazy Language?

Your colleague argues this language is used for “evil purposes or simply to pump up the associated stocks.”

While stock pumping and marketing hype are absolutely driving a lot of this (tech companies love to imply their software is close to “God-like” Artificial General Intelligence to attract investors), there is a third, more innocent reason we use these terms: Metaphorical Convenience.

Humans are biologically wired to anthropomorphize things. We say our car “doesn’t want to start” on a cold morning, or that our computer is “tired.” Saying “The LLM predicted the next token based on a probability distribution conditioned on the prompt context” is a mouthful. Saying “The AI thinks you mean X” is lazy, shorthand slang that unfortunately confuses the general public.

Summary

Your colleague’s response is a refreshing dose of engineering reality. Strip away the sleek user interfaces and the marketing buzzwords, and AI is precisely what they said it is: an incredibly massive, ultra-complex math equation calculating an output based on an input. It doesn’t know what it’s saying, and it doesn’t care.

Surely this is how we should be educating the public about “Artificial Intelligence”?

How many people are turning to an A.I. Chatbot and asking it for relationship advice? Medical diagnosis? Marketing help? How to run a business? And how many of those KNOW that all that is happening is a mathematical prediction model NOT something with feelings, empathy, life experiences and emotions giving them “advice“?

(Any yes, I realise that I have used EXACTLY what I am warning about, to write the responses and put them in full into this blog – but isn’t it amazing how we “believe” or “trust” what they produce, when they align with our own opinions!)

Categories +

Get in touch

If you have any questions or if your website needs optimising to reach your target audience, contact me on the below:

Blog