I have been musing this on and off since R. Daneel Olivaw. It is tempting to wonder if R. Daneel Olivaw was created by humans, or did it (He?) spontaneously arise without intelligent living intervention?

It seems to me that the creation of Artificial Intelligence goes against entropy. Not the execution of the intelligent computations, the electrons whirling, the gears turning, the power being consumed, friction wearing things down, chips burning out, but the design/creation of the device in the first place? In order to have any logical device, there must be extreme order and complexity. Memory is highly organized, interconnected, symmetrical, and in its general specifics entirely the opposite of what one would expect as a result of entropy. The same for logic gates, and the hierarchical and algorithmic interconnection of them. Somehow, they would have had to become organized to begin with. I can not envision, even as a result of infinite randomness, the development of an intelligent device based only on the interaction of the laws of physics applied to inanimate matter, let alone its evolution into even more complex intelligent devices. There would, it seems to me, have to be some design of the system in the first place, some planning of the organization, connections, and process.

Thus, it seems to me that the very nature of entropy would deny the possibility that complex organized intelligent devices could ever arise based on the processes of the physical world alone.

If there was never any life in the universe, could AI devices ever be created, or does the existence of AI always imply that there originally was some form of intelligent life to begin with?

Is there any possible, or plausible, way that in my universe, without hand waving, R. Daneel Olivaw type construct could ever occur without the intervention of intelligent life? I am not concerned about it being a robot, I would be happy with a smart information processing rock that developed mobility through the application of intelligence that it gained spontaneously.


There appears to be an issue with the word 'Artificial Intelligence'. Let me clarify it as meaning 'non-living intelligence', that is intelligence that is not nor ever was 'living', as opposed to intelligence as we know it, that is a result of or based on life.


I absolutely never imagined having to define 'living' as in 'animate' vs 'not living' as in 'inanimate'. A 'living body' is something that we all classify as a 'not dead' body. I am not about to go into the details as to how we determine that a mass of elements, molecules, gasses, protoplasma, tissues, muscles, bones, and such that is at one moment called a 'living body' and at the next moment called a 'dead body'.


I am aghast at how many people do NOT know what Artificial Intelligence is. There are a LOT of misconceptions about it. it does NOT mean 'made by humans', it means 'not human (or life) intelligence'. If it is not 'real' human (or life) intelligence, it is Artificial Intelligence. The definition does NOT indicate how the artificial intelligence device was made, it means simply that it is intelligence that mimics human (life) intelligence but is not in a human (living being). It is a standard definition. Every google source I tried gave the same equivalent definition. Intelligence that mimics human intelligence. Absolutely never any reference to how the device that does the mimicking is made. I am using a standard term with a standard definition.


The most intriguing answer so far is that weather systems are indeed a form of artificial (non-human or non-living) intelligence that is 'natural' (for those arguing the definition), not human-made. They have the obvious ingredients of data (temperature, pressure, air volume, wind speed, etc.), input, processing, output. You have reprogrammability (the landscape changes). You have logic (if-then) operations. You have comparisons and resulting decisions. You have repeatability. You have a deterministic, algorithmic model. It does not 'think'. All the criteria for an AI. Now, could it transmute into a different application? Could it generalize? By what creative possible or plausible mechanism? Are there other such examples of 'machine (non-living) intelligence', created without any intervention from living thinking beings?

  • $\begingroup$ Comments are not for extended discussion; this conversation has been moved to chat. $\endgroup$ – L.Dutch Mar 12 at 6:07
  • 1
    $\begingroup$ "Absolutely never any reference to how the device that does the mimicking is made": the keywords here are device and made. I've been working in IT for a quarter of a century and I have never ever come across an understanding of AI as anything else than intelligence exhibited by computing machinery. Note "machinery". The machines in question can indeed be studied as abstract mathematical objects, but any physical realization must indeed be "artificial" if their intelligence is to be considered "artificial" intelligence. $\endgroup$ – AlexP Mar 12 at 9:16
  • $\begingroup$ @AlexP Which is exactly the essence of the question. Can only humans make machines? Or can machines be made without life? The assumption is that only humans (or aliens equivalent to humans) can make machines. I am questioning this assumption. $\endgroup$ – Justin Thyme the Second Mar 12 at 14:52
  • $\begingroup$ "Can machines be made without life?": Of course they can: machines can make machines. We have now machines which make machines; both in the physical realm (robots in factories) and in the abstract realm (programs which write programs). The question is who made the machines which made the machines, ..., which made the machines; and there are only two answers: either God (who was neither "made" nor "created") or some naturally evolved life forms, or if you prefer, machines which exhibit the fundamental attributes of life (metabolism, reaction to stimuli, reproduction). $\endgroup$ – AlexP Mar 12 at 15:15
  • 1
    $\begingroup$ @AlexP WHO made the first machine? Obvious bias. WHAT made the first machine? Still bias. HOW was the first machine made? The point of the question. Other options, that intelligence is a fundamental property, or quantum field, of the universe. Or that Boltzman Robots spontaneously happen, instead of Boltzman Brains. Or that 'life' is not about metabolism, reproduction, cellular function, but can be something we do not even recognize as life. All answers that have been posited. But all definitely outside the 'comfort zone' of many around here. THESE are the answers I am looking for. $\endgroup$ – Justin Thyme the Second Mar 12 at 15:43

Thus, it seems to me that the very nature of entropy would deny the possibility that complex organized intelligent devices could ever arise based on the processes of the physical world alone.

If you accept that we humans are intelligent, you are also accepting that complex organized intelligence can arise based on the process of physical world alone, falsifying your statement. We are nothing more than carbon based devices.

Could Artificial Intelligence ever arise without life?

If with artificial you mean "not directly made by the natural world", then you cannot have anything artificial without the mediating intervention of some intelligent life form which is able to manipulate and create artificial things. Whatever is generated by the processes of physical world alone is, necessarily, natural.

  • $\begingroup$ But we are a LIVING intelligent being. I want a non-living, never-has-been-living intelligence. $\endgroup$ – Justin Thyme the Second Mar 11 at 19:02
  • $\begingroup$ Your first point belies your second point. But I am not in this to argue if human intelligence was designed by some creator, only that 'living' is not the same as 'inanimate'. $\endgroup$ – Justin Thyme the Second Mar 11 at 19:23


Your answer is in your question. The word artificial means "Created by Humans" or other life forms. If it occurred "Naturally" it would not be called "Artificial" It would just be another form of actual life. Your question about entropy is a good one, and is actually one of the most common talking points regarding the origin of life itself and whether a creator is necessary. But that's a question that really doesn't belong in world-building.


There is no universal definition of life, and by extension no universal definition of intelligent life. If you did find a "Machine" (by your definition) that had "Artificial Intelligence" (also by your definition), that was not created by "Intelligent Life" (by your definition), then people would probably redefine life to include such a "machine", then stop calling it a machine, and would not classify it as "artificial intelligence" anymore, but rather another form of intelligent life.

Consider, for example, that an alien race which developed only "DNA" computers (although their form of life might be completely different from ours) were to discover a human (a life form based on how their computers work) and declared humans to be examples of artificial intelligence, because to them we are machines. Would they be right, or would they just have discovered another form of intelligent life?

  • $\begingroup$ All life occurs 'naturally'. Cells divide naturally. Animals give birth naturally. But life also has artificial insemination. Both 'natural' and 'artificial' insemination is something 'created by humans'. The term 'naturally' has nothing to do with animate vs inanimate. $\endgroup$ – Justin Thyme the Second Mar 11 at 19:32
  • $\begingroup$ 'In computer science, artificial intelligence (AI), sometimes called machine intelligence, is intelligence demonstrated by machines, in contrast to the natural intelligence displayed by humans and other animals. Computer science defines AI research as the study of "intelligent agents": any device that perceives its environment and takes actions that maximize its chance of successfully achieving its goals.' from en.wikipedia.org/wiki/Artificial_intelligence I have looked up numerrous definitions, and not ONE says how the machine was made or who made it. $\endgroup$ – Justin Thyme the Second Mar 12 at 2:30
  • $\begingroup$ Artificial Intelligence does NOT mean 'Created by humans or other life forms'. It means 'not human intelligence'. $\endgroup$ – Justin Thyme the Second Mar 12 at 2:31
  • 3
    $\begingroup$ That is a very unusual definition, I doubt anyone else would call aliens artificial intelligence. The word artificial means created or made by humans, or in this context also aliens. $\endgroup$ – Whitecold Mar 12 at 8:57
  • $\begingroup$ @Whitecold I can cite definition after definition, from IBM, universities, think tanks, google, auto makers (for self-driving cars) that all refer to human intelligence. It is NOT an unusual definition, it is the one used by almost everyone in the field. Get over you hangup with your definition of 'artificial intelligence'. It is far too limiting. $\endgroup$ – Justin Thyme the Second Mar 12 at 14:57

Semantics aside over definitions of 'artificial', I think what you are looking for is a theoretical idea called Boltzmann Brains.

They're (presumably very rare) thinking structures that spontaneously coalesce out of random fluctuations of matter. No reason they're incompatible with the concept of entropy.

  • $\begingroup$ Comments are not for extended discussion; this conversation has been moved to chat. $\endgroup$ – HDE 226868 Mar 11 at 23:59

One of the more interesting emergent new definitions of life orbits this very idea - that life is in fact an emergent property of aggregate matter which locally defies entropy, and does so specifically in a fractally-organised manner... that this may in fact be simply an emergent property of matter aggregates in our particular universe.

If this proposition is true, it implies not only that life is not necessarily unique to our corner of the universe, but also that life may take wildly divergent forms and paths, and that in the end the correct working definition may have nothing whatsoever to do with things like metabolising chemistry, reproduction and other such metrics with which we have been obsessed, but rather with the net effect of a given class of entities, artifacts or events on their local space.

Under that definition, it may well be that certain types of organic-but-currently-seen-as-non-living chemistry which has been observed in meteor remnants, comets et al in fact do represent life, just of an order we don't current recognise as such.

Given that background, it's worth pointing out that we may well be literally incapable of conceiving of truly alien life, or recognising it when / if we observe / encounter it; I see no objection to paths of evolution arising from non-human life which could, over time, lead to the development of AI in some form: it would be intrinsically unrecognisable to us, and for all we know, we've already encountered it (say via FRBs) and have not the slightest idea what we're looking at.

To give a corollary posit: we cannot currently "decode" the content of a signal from one of our own cerebral neurons, and were we "looking at" the signal it generates as a piece of contextless radio data, we'd have no clue what it was, much less how to "read' it - or that it even could be containing data. Given this, if in fact all the nearby stellar nebulæ are part of a very large, slowly-communicating inorganic AI "brain", and the recently much-ballyhooed observance of FRBs is our first inkling of a vast sentient signalling network in which we're embedded, how could we conceivably tell this from the observable data?

  • $\begingroup$ A truly refreshing answer. It posits, in some way, that intelligence itself locally defies entropy, and that something that locally defies entropy must be intelligent. Essentially, there is only one life on earth, because anything that is living got its life in succession from that one first living thing. That is, life does not begin at 'birth', birth is a continuation of the life that was in the single gamete cells which are art of the living parent organism which came into being because a living cell was passed to it form its parent organism and so on back to the very first organism. $\endgroup$ – Justin Thyme the Second Mar 11 at 20:38
  • $\begingroup$ But what it does NOT address is that. in every subsequent living organism that this one living cell propagated, things became more and more intelligent, not less and less, as entropy would suggest. That is, all life on earth is continuous with, and part of, that original living cell. Life is not discrete, it does not begin and end, it is continuous. All life is part of that original life, and that entire process is one of defying entropy at a greater and greater scale. $\endgroup$ – Justin Thyme the Second Mar 11 at 20:39
  • 1
    $\begingroup$ And thus, the question: If intelligence, as a completely separate issue to what we call life on earth, has the ability to propagate exponentially (and at some high power exponent) throughout the universe, can we still use the term 'locally' in 'locally defies entropy'? At what point does exponential growth throughout the universe become 'universally defies entropy'? If matter itself can become intelligent, and this intelligence spreads exponentially, would not entropy become 'balanced' between entropy/non-entropy? A truly 'living, intelligent universe'? $\endgroup$ – Justin Thyme the Second Mar 11 at 21:03
  • 3
    $\begingroup$ @JustinThymetheSecond The evolution of life does not defy entropy in any way. This is an idea generally espoused by creationists as an argument against evolution but is not considered a valid criticism by the scientific community. Entropy is a very difficult concept and this is exploited by those who wish to forward certain beliefs. If you are interested in arguments debunking this claim I can direct you to (TalkOrigins)[talkorigins.org/faqs/thermo.html]. However, this comment section is not a good place for an in-depth debate or discussion on thermodynamics and evolution. $\endgroup$ – Mike Nichols Mar 11 at 21:29
  • 1
    $\begingroup$ "Entropy is a very difficult concept" ~ Oh FFS! ~ It's just the 2nd law of thermodynamics, not some incomprehensibly nebulous philosophy ~ there is absolutely nothing even slightly nebulous or difficult to grasp about it, but it does have p#ss all to do with evolution so @MikeNichols [+] for that much. $\endgroup$ – Pelinore Mar 11 at 22:23

The answers provided so far are in fact an example of a collective intelligence, which while it is made up of live creatures, is not alive as an individual organism and therefore cannot really be considered 'alive'. It is interesting that every answer provided so far is correct, albeit from different points of view;

1) Dutch and Mathaddict have both pointed out the definitional problem with the word Artificial
2) Ynneadwraith has (quite rightly) invoked Boltzmann Brains, and
3) Gerard has pointed out the relationship between entropy and the ability to store order.

Putting this all together, including the edits that you've made to your question, then yes, I believe that non-living intelligence occurs around us all the time. The reason for this is often that life powers it, but not always. But to clarify the question first, it might be useful to define intelligence.

Intelligence is the ability to identify or recognise complex patterns. That said, that's not how we measure it; we can't see inside a mind or computer model and know what patterns it is or isn't capable of recognising or identifying, so we use a slight variation to the definition to measure; we measure intelligence by how an intelligence reacts to complex patterns.

If you take this as an empirical measure of intelligence, then arguably the earth's atmosphere is intelligent. If wind isn't present, electrical storms create fires, which create smoke, which lets in less light, cooling the area down. Patterns like el nino are generated by warmer waters in the Pacific. The atmosphere reacts in consistent ways to complex phenomena or patterns. The atmosphere is not alive, has no free will, and certainly isn't conscious, but it's reacting consistently to complex patterns. This (by the definition above) is a form of intelligence.

I mentioned collective intelligences before; take a look at ants. Ants do not have the brain capacity for reason or architecture or engineering, yet collectively they build and run small cities together, showing tremendous organisation. There has been a great deal of academic discussion on the collective intelligence of ants, but the collective is not alive per se; it is made up of a myriad of living things, which is very different.

Of course, one of the most insightful comments from above is Dutch who refers to us as 'carbon machines'. The reason why this is important is because there is a great debate (and I'm really hoping I'm not opening up another can of worms here) around whether or not the universe is deterministic or not. This ties into the debate about whether or not we can truly have free will.

Ultimately, if the universe IS deterministic, then what that means is that if you know the algorithm for the universe AND you know the state of the universe at any given point, you can calculate the state of the universe at any other point in time, past or future. That would mean that free will can't exist.

If that's the case, then a non-living intelligence is certainly possible as our own sense of free will has to be a lie under those conditions, so it would mean intelligence cannot be a function of life; it's a function of the complexity of the universe itself.

The thing is, the universe acts in a mostly deterministic way (I say mostly because I can only speak for what we've measured) at the macro scale. We know that relativity works as a model for predicting the future and we base orbital mechanics and many high cost space missions on those rules and get it right. But Quantum Mechanics conforms to the scientific models we currently have of how things work on a very small scale. Ultimately we don't know how much of an impact that has on the fabric of the universe yet as this is a very big question and at the macro scale at least, minor variations seem to cancel each other out very consistently.

The point being, that if you think that intelligence (not liveness, consciousness, sentience, or sapience) is a function of something unique to a living brain, because of its ability for localised reversal of entropy, and you believe that this isn't possible outside of a living brain, then your answer is no. But, if you believe that intelligence doesn't require localised entropy reversal, or that localised entropy reversal can occur outside of a living brain, then your answer is yes.

Ultimately the answer to your question is governed by the model of the universe you hold to. It's also worth pointing out that at some level, the universe itself might just be a form of non-living intelligence.

In any case, it is also important to note that this is a question that has kept many scientists and philosophers awake at nights for millennia; which is another way of saying I'm not so arrogant to think I have the complete answer in this post.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.