Simulating is not feeling: the invisible limit of artificial intelligence

Think about it for a moment.
You ask a question in front of a screen and, within seconds, a response appears that seems perfect. Well structured, nuanced, even with a touch of humility when it self-corrects. Sometimes it seems to hesitate; other times it seems to reflect. And then that strange feeling surfaces: is it actually understanding me?

This is not a technical doubt — it is something more instinctive. When a system responds convincingly, our brain fills in the gaps and assumes there is someone on the other side. But here is the key: the fact that something behaves as if it understands does not mean that it does.

Subtle differences
It is easy to confuse what we observe with what is actually happening. Artificial intelligence can solve complex problems, recognise patterns and generate highly coherent responses. That is undeniable. But executing calculations or drafting a correct piece of text is not the same as experiencing it. An algorithm can give the right answer without having any idea what it means. A calculator performs square roots at absurd speed, yet it has no concept of number and feels no satisfaction at reaching the result.

The same applies to modern language systems, such as the chatbots that imitate human conversation. They identify patterns and generate convincing responses, but there is no inner experience behind what they produce. No emotion, no intention, no real understanding — only highly polished outputs that appear "intelligent".

The convincing illusion
Why does it feel so human? Perhaps because we are wired to detect intention everywhere. Evolutionarily, assuming that something has purpose kept us safe from invisible dangers. Today, that same sensitivity leads us to assume there is an "interior" in a machine, when we are simply observing patterns that work.

If a system maintains coherence, adjusts its tone and corrects errors, our mind interprets this as evidence of consciousness. What we perceive is behaviour; what we assume is understanding. And that is where the illusion lives.

The invisible boundary
The true limit is not how fast it calculates or how much memory it holds. It lies in real experience: whether there is a "within" where something actually feels what it processes. From the outside, we might think we have already crossed that line. But from the inside — if there is an inside — the difference would be radical. Consciousness cannot be measured with sensors or algorithms. That is precisely what makes it invisible.

What this means for us
Believing that a machine understands when it only simulates can lead to real mistakes. We inflate expectations, delegate decisions to systems that do not grasp consequences, and let ourselves be carried along by narratives that mix fascination with fear. The real risk is not that AI advances too fast — it is that our perception advances faster than reality.

Exploring the possibility of artificial consciousness is not absurd. There is legitimate research in neuroscience and robotics seeking to understand how experience is generated. In fact, part of the future work at IA Robots will be to investigate how to approach that boundary without losing perspective: how to design systems that can learn autonomously and, perhaps someday, come close to a conscious experience — without confusing simulation with reality.

For now, what we have in front of us is not experience — it is simulation. A chat that responds perfectly does not "feel" or "understand", however much it may seem to. Recognising this difference is essential if we are not to lose our bearings while we keep building and experimenting.

And perhaps the most interesting question is not only whether a machine will ever feel, but how we will react when we face that mirror of our own intelligence — one that today merely imitates us, but that in the future might begin to experience the world for itself.