(Non-naturalist views could also serve this purpose. One might hold, for example, that God installs conscious souls by miracle. But most mainstream philosophers and scientists prefer explanations in terms of purely natural phenomena.)
Searching for a Biological Property That Is Necessary for Consciousness
Biological naturalist arguments against AI consciousness can be expressed in two steps:(1.) Property X is a biological property, of the sort that cannot be instantiated in standard (non-living) AI systems.
(2.) Property X is necessary for consciousness.
Different versions of biological naturalism propose different candidates for X.
Peter Godfrey-Smith for example, appeals to metabolism. For this appeal to work as an argument against AI consciousness, two things have to be true. The relevant sort of metabolism must not be instantiable in standard AI systems, and a system without metabolism in this sense must necessarily lack consciousness. Given the first condition, drawing power from a solar panel can't count as "metabolic". Godfrey-Smith suggests that molecular-level spontaneous motion is crucial to metabolism in the relevant sense. The second condition requires that metabolism in this specific sense is necessary for consciousness. Godfrey-Smith suggests it might be necessary, but he offers little positive argument. The suggestion remains an intriguing possibility rather than a fully developed position.
In an already-influential article published last month in Behavioral and Brain Sciences, Anil Seth proposes autopoiesis as the X. Autopoiesis -- a concept first developed in a 1972 book by Humberto Maturana and Francisco Varela -- occurs when a system continuously regenerates its own material components, actively maintaining a boundary between itself and its surroundings (Seth 2026, sec 4.1). For autopoiesis to serve as X, standard AI systems must be incapable of it, and consciousness must require it.
Neither claim is obvious. Regarding the first, it seems that we can, with a bit of creative engineering, design an AI system that regenerates its own material components and enforces a boundary between itself and its surroundings. Imagine a solar-powered robot that seeks out solar charge. It has redundant parts that it monitors for defects and which it can replace and rebuild by ordering shipments of snap-together components. For more details, see this post and my commentary on Seth's article.
Regarding the second claim, even if we grant that autopoiesis is limited to biological systems, it's not clear why consciousness should require it. Seth appeals to predictive processing and Friston's free energy principle, but these appeals are highly speculative. They do not establish the necessity of autopoiesis for consciousness, as several of the other commentaries emphasize.
There's substantial plausibility in the thought that consciousness requires using energy and some kind of stability over time. But even granting this, it remains unclear why consciousness should require metabolism or autopoiesis understood narrowly enough that no ordinary AI system could instantiate them.
Some of the replies to Seth's article suggest other candidates for X. For example, Cao, Gottlieb, and Moore propose evaluative states that are internalized and integrated. Again, the view is potentially attractive, but both premises are highly debatable.
First, it's not clear that a sufficiently sophisticated AI system couldn't meet Cao and colleagues' condition. It might have an optimization process that corrects deviations from a goal state, plus explicit representations of these processes, plus the capacity to integrate many factors in achieving one or more goals. This seems especially plausible for autonomous embodied robots who must negotiate complex situations in the wild. Second, it's not clear why consciousness requires valuing rather than simply registering. It might! But this is the type of proposed requirement for consciousness that I criticize in my recent book AI and Consciousness: broadly plausible, but difficult to defend rigorously.
Biologicity as a Matter of Degree?
In his reply to my commentary (section 5.2 here), Seth notes that my hypothetical robot depends on the environment for ready-made components. But this is arguably only a difference of degree from the human case, not a difference in kind. We humans don't manufacture all of our components from the atomic level up. We obtain some amino acids and vitamins from the environment because we can't produce them ourselves. And my robot does a little self-manufacturing: It snaps together components of a replacement leg.
Seth writes:
On the other hand, if a robot could indeed manufacture its own components from generic substrates under energetic constraints, then it might be truly said to be autopoietic, but it would also be considerably closer to a living system than any existing computer or robot.
So I think we should imagine a spectrum of biologicity, not a sharp divide between the biological and non-biological. If autopoiesis marks the biological, we can conceive of systems ranging from those that self-maintain in only the most limited sense, to those that do somewhat more, to those that self-construct at many levels at once, including their chemical components. AI systems might advance partway along this spectrum with no radical change in their underlying technology. But I'd agree with Seth that systems that manufacture even their own chemical components might merit the label "biological" -- perhaps as instances of artificial biology.
Maybe we can call systems in the middle of this range semi-biological. One might imagine semi-biologicity for other processes too: metabolism that somewhat but only somewhat resembles the kind highlighted by Godfrey-Smith, embodied goal integration of the kind highlighted by Cao and colleagues. Rather than thinking of the "biological" in a sharp-edged way, we might imagine ranges of artificial systems that have relatively more or less of various biology-like properties.
Of course, none of this addresses the second premise. Why should autopoiesis, or any other such biology-like property, be essential to consciousness? If we're inclined to reject near-term AI consciousness, requiring at least mid-grade autopoiesis or biologicity might yield intuitively the right results. But a rigorous defense of that requirement is still lacking.
ETA, Oct 3: Godfrey-Smith replies on X that his aim in his 2016 paper was more like "here is what it is, in us and other animals, that might enable us to get a grip on it". So it would be an interpretative reach to rely on that paper for a version of Premise 2 above, but of course one might also be inspired to consider whether Premise 2 would be true. He also points to a more recent paper of his that emphasizes the importance of large scale dynamical patterns in nervous systems (which would be difficult, but perhaps not impossible, to implement on current hardware).
I was going to write some neuroscientist thoughts on this, but then I realized I wrote all those thoughts a couple months ago :) https://schwitzsplinters.blogspot.com/2026/06/do-computers-have-wrong-substrate-for.html
ReplyDeleteMight there be another level of explanation here: comparing biological and artificial systems through the engineering-level causal organization of information processing?
ReplyDeleteConsider a system whose continued existence depends on conditions such as energy and functional integrity. Differences concerning those conditions influence competing candidates. When one incompatible continuation is realized, the others lose their standing to determine that continuation. Its consequences change the same system, and subsequent perception and selection begin from those changed conditions.
Remembering the unselected path does not mean it was taken; forgetting the selected path does not undo its consequences. History therefore operates as present conditions shaping what comes next. No additional observer is needed to read it: those conditions influencing subsequent processing perform that role.
Computers also have a single physical history and feedback. The further question is whether selection changes their own viability conditions and reorganizes their possibilities as a whole. In output selection, the winner determines the next output; in this proposed structure, its consequences determine the next conditions of the same continuing system.
This suggests a concrete artificial design. A robot might choose between crossing a dangerous bridge to recharge and taking an energy-consuming detour. If crossing damages its leg, subsequent perception and selection must proceed from that damage and remaining energy. Retrieving the detour plan does not erase the consequences. Your self-repairing robot could implement this linkage.
The system forming perception, bearing consequences, and continuing from them can then be traced as one causal unit. Its correspondence with one subject’s present and history supports an inference that this unit is itself the subject. An additional subject or ownership mechanism would need to explain an identifiable missing role.
This does not establish the mechanism of felt quality. Yet perceptual information participates in forming this system’s actual present and continuation. The remaining question is why that perceptual event is felt, and why it feels red.
This approach does not require biological materials. It asks what causal roles metabolism or autopoiesis perform, and whether other implementations can perform them. Could candidate X therefore be investigated as a causal organization rather than solely as a named property?
(Please allow me an additional comment.)
ReplyDelete1. A hypothesis of first-person subjectivity arising from an engineering-level causal structure
Candidates concerning a system’s own viability compete, and one is realized. The same system bears the consequences, and subsequent perception, evaluation, and candidate formation begin from its changed conditions. History operates not merely as a record, but as conditions organizing subsequent processing.
This role is what the paper calls the “internal reference point.” No additional observer is needed to read the history. If perceptual formation also changes the same system, and those changes feed back into formation, the conditions of evaluation themselves are updated.
If the unit forming perception, bearing consequences, and continuing from them can be traced through the same causal linkage, might the system realizing that linkage itself be the subject with a first-person present and history? This is a strong inference about subject structure. The mechanism of felt quality remains unknown, but perceptual information participates in events forming the system’s present. This makes the next question more specific: through what mechanism does this subject feel red?
2. Could brains amplify this causal structure?
The same criterion applies to single-celled organisms. We do not yet know what differences in the emergence or intensity of consciousness follow from differences in this structure, or whether those differences form a continuum.
Brains might nevertheless amplify the linkage by bringing sensory information, bodily states, memory, and prediction into candidate competition concerning the whole body, and implementing closure and consequence-driven updating over broader scopes, at greater speeds, and across longer timescales. This predicts shared functional structures addressing the same causal task across species.
Despite differences in neural circuitry and anatomical organization, research has examined candidate competition, enhancement of selected activity, and suppression of unselected activity in different species. These findings provide evidence relevant to shared causal roles, beyond anatomical resemblance alone. How those processes connect to bodily viability and continuation from consequences still requires further investigation.
Applying the same framework to biological and artificial systems allows comparison of its realization in single-celled organisms, its amplification by brains, and its artificial implementation. Might this provide a consistent engineering perspective on both subject structure and the evolutionary role of brains?
(Final comment)
ReplyDeleteIf self-generation is to be considered a condition for consciousness, should we not also examine the relationship between the process that maintains the self and the process through which an intention becomes consciously accessible to that system? This is a shared question when comparing biological organisms and AI.
Here, the time lag between preceding brain activity and conscious awareness of intention in Libet’s experiments offers a clue for thinking about causal structure.
If we call the role through which a system’s history organizes its present perception, evaluation, and candidate formation an “internal reference point,” this is not an observer reading information, but a function that is updated within the same system. Candidates form and compete under these conditions, and that process further changes the system’s current conditions.
If conscious awareness of a particular intention emerges at a certain stage of this formation process, a time lag between the beginning of the process and that awareness is natural. We can understand both as parts of the same subject’s formation process, without placing the consciously aware subject outside the preceding activity.
This does not, however, directly identify the onset of the readiness potential with the beginning of candidate formation, or the reported time of intention with the moment when the internal reference point is updated. Which processes give rise to conscious awareness of intention remains a question for investigation.
Alongside the question of whether self-repair or self-generation can be artificially implemented, we should also ask how intention-forming processes within such an implementation relate to first-person awareness. By comparing this temporal causal structure as well, might we examine consciousness in biological organisms and AI using a common framework?
I think there has been movement and work towards biological naturalism for some time. My elder brother was investigating this before either of us knew the term. (since around 1997).He has long been interested in the implications/effects of historical cosmology. I too have a great deal of curiosity and interest and will ensure he knows of this post.
ReplyDeleteAdditionally: "The Human Potential Movement (HPM) emerged in the 1960s around the belief that humans possess vast, untapped creative psychological. capabilities. While traditional Metaphysics asks, "What is the nature of reality?", Meta-Metaphysics shifts the lens to ask, "How do our structures of thought, language, and consciousness construct our understanding of reality itself?""...
DeleteWithout biological consciousness we would not know we are here...Human consciousness is next to...hereness and its' meanings...
ReplyDeleteMy brother grew among this epiphany. I was several years younger.
ReplyDelete