This week, I again watched two reports on artificial intelligence, and this time my theme was: does artificial intelligence have consciousness, or can it ever develop it? That is, of course, a chilling question, and first, I watched a report or an interview with Geoffrey Hinton, who is a Nobel Prize winner in 2024, a specialist in the field of neural networks, and for him, it is absolutely clear: artificial intelligence already has consciousness. He has no doubt about it.

At one point, the interviewer asks him where he gets that certainty from, what it is based on, and he actually responds very briefly: Well, yes, we have conducted experiments with neural networks. When you remove a part of them and replace it with a nanoparticle with the same properties, the consciousness remains intact.

I don’t know if that is a scientific explanation because I immediately have questions. I think—well, in the first place, I assume this was not tried on humans, but secondly, if this were to be tried on humans or eventually applied, then you know that when a piece of brain tissue is dead or so diseased that it no longer functions, it is, if possible, taken over by other tissue. If the damage is very extensive, of course, that does not happen, and then you could imagine that it would be wonderful if you could insert a nanoparticle that replaces that function.

But to then say that this is proof that such a nanoparticle also has consciousness—I do not consider that a scientific conclusion. Because there are so many other possibilities by which consciousness could be maintained. So, this made me somewhat uneasy. Such a highly learned man, who is engaged in scientific research at such a level that he has received a Nobel Prize—perhaps it was the interview, but I was not convinced at all. It would be difficult to convince me, but well, that is another matter.

I do want to try to follow the reasoning as open-mindedly as possible, and here I simply could not continue. Well, you know, once you have watched such interviews, you are also shown others, and then there was an interview with a very old Roger Penrose, who, as I mentioned earlier, wrote this book! First edition in 1990, so you could say, well, that is long outdated.

But if we want to gain any insight into what artificial intelligence actually is, then we must also go back in time—to the time when it was still somewhat comprehensible what was happening. And I already found that sympathetic back then, because he works in an extremely scientific manner but, at the same time, asks very concrete questions.

And in this interview with this old Penrose—and I don’t know from what year this interview is, but in any case, he is old—he tells the interviewer that, as a mathematician, as a scientist, he has had a moment in his life, or perhaps a series of moments and a shift in his awareness, that understanding is not computable.

I had to listen several times: what is he saying? Because he repeats it multiple times—computable, it is not computable. And at a certain point, he says: It will never be possible to encapsulate human understanding in an algorithm because it is not computable.

And then, of course, you become very curious about whether he has a vision of what human understanding actually is. But first, I want to read a passage from that book, *The Emperor’s New Mind*, from 1990. I have previously read something from it here, but that was about a man who entered a hamburger restaurant. Certain events take place, and then the computer must determine what the final conclusion of those stories is.

There was a certain John Searle who believed that the fact that a computer provides the correct answers does not at all mean that it also knows what it is doing, and he wanted to prove that. He tried to demonstrate this through an experimental setup, which I want to read out:

The question that we must consider is whether this kind of success actually indicates any genuine understanding on the part of the computer or, perhaps, on the part of the program itself. Searle’s argument that it does not is to invoke his concept of a “Chinese room’. He envisages first of all, that the stories are to be told in Chinese rather than English surely an inessential change- and that all the operations of the computer’s algorithm for this particular exercise are supplied (in English) as a set of instructions for manipulating counters with Chinese symbols on them. Searle imagines himself doing all the manipulations inside a locked room. ‘ So he is sitting in that room and performing those operations as if he were a computer.

The sequences of symbols representing the stories, and then the questions, are fed into the room through some small slot. No other information whatever is allowed in from the outside. Finally, when all the manipulations are complete, ‘ “—so he does exactly what the algorithm does— ‘the resulting sequence is fed out again through the slot. Since all these manipulations are simply carrying out the algorithm of Schank’s program, it must turn out that this final resulting sequence is simply the Chinese for ‘yes’ or ‘no’, as the case may be, giving the correct answer to the original question in Chinese about a story in Chinese.’ You have to let this sink in for a moment—what a remarkable situation this is. Because there is a person in a room who receives a story in an unfamiliar language, so he cannot understand it, but he has been given a program that allows him, step by step, to determine whether the answer should be yes or no, and he ultimately does that, and it goes back outside through the mail slot.  ‘Now Searle makes it quite clear that he doesn’t understand a word of Chinese, so he would not have the faintest idea what the stories are about. Nevertheless, by correctly carrying out the series of operations which constitute Schank’s algorithm (the instructions for this algorithm having been given to him in English) he would be able to do as well as a Chinese person who would indeed understand the stories.

Searle’s point- and I think it is quite a powerful one is that the mere carying out of a successful algorithm does not in itself imply that any understanding has taken place. The (imagined) Searle, locked in his Chinese room, would not understand a single word of any of the stories!’

It is an almost ludicrous experiment, and you can hardly escape this conclusion: that if you do not understand anything and simply follow a sequence of steps, you still manage to arrive at the correct answer.

There has been a lot of criticism of this, which I will spare you, but let’s stick to what emerges in this passage and to Penrose’s statement that, in his life—this is, I believe, truly an inner experience, that is how it comes across—he realized: when I engage in scientific work and I understand what I am doing, then that understanding, that process, is something that is completely inconsistent with everything we know about what can be computed.

That is what he experienced, and then I think: why did he not find Rudolf Steiner? Because what happens? He naturally remains within natural science, which is very understandable—he does not turn to spiritual science. He remains within natural science, even though he says “understanding is not computable” and even at one point more or less suggests that it is not a material phenomenon.

Then I would say: seek spiritual science! But I believe that for someone so deeply immersed in natural science, that is a difficult step.

In any case, he apparently did not take it, because he ultimately arrives at—not really a conclusion, but rather a pointer towards quantum physics, suspecting that the mystery of consciousness might be found there.

Well, let’s leave that with him—that is his conclusion. But what lies in between—the step from the computable nature of intelligence to the non-computable nature of understanding—that is an enormous step, one that should make you want to run jubilantly through nature, joyfully realizing that human intelligence truly cannot be calculated.

They can do whatever they want—they can refine artificial intelligence further and further, they can develop processes to make it self-learning so that humans are no longer needed at all—but what they cannot do is make artificial intelligence function the same way as human intelligence through an algorithm.

But of course, there are many opponents. This Geoffrey Hinton now has a completely different vision, truly believing that it has already happened.

And my view is that true certainty can only arise if a moment like Roger Penrose’s occurs in your life—when, while engaged in understanding, you have an experience that understanding is truly something suprasensory, that your brain alone is not enough.

That is ultimately what it means: understanding does not happen with the brain. And a computer does not even have a brain—but it has simulated neural networks. That can be simulated, but the process of understanding, so beautifully described in the Chinese room, is completely absent while still arriving at the right answer.

That process of understanding is a human miracle that far surpasses natural science. To comprehend this, you need spiritual science.

And I am convinced—this is what I wanted to say—that if enough people inwardly recognize what human understanding truly is, then artificial intelligence may advance tremendously, but it will never prevail.

I would like to continue this discussion next time.