On artificial and natural intelligence
From fictitious reading lists to real rivers
In electric dreams
This week, everybody seems to be talking about the fact that AI can “hallucinate”.
It started with my husband sharing this article, which highlights that “as [AI] models grew in complexity, so did the nature of their errors. Today’s advanced language models, designed to 'reason’ through problems, can produce lengthy, convincing explanations that are entirely untethered from reality.”
Not long afterwards, I read that Elon Musk’s chatbot Grok had been caught promoting false narratives of “white genocide” in South Africa - even in response to totally unrelated queries. Then we heard that AI had been used by a newspaper to create a summer reading list which contained several recommendations of books that didn’t exist.
Meanwhile – on the other hand – an acquaintance shared how she had been using ChatGPT as a therapeutic tool, first training it in the kind of support she wished to receive, and then calling upon it for advice and affirmation at difficult moments. She described it as “horrifyingly good”, saying it helped her to be with her body and listen to her emotions.
Can a computer think? Can I?
To shed light on this – as with many other vexed issues in my life – I was drawn to look again at Gregory Bateson’s Steps to an Ecology of Mind (1972), one of my favourite books ever. Being among the pioneers of the new field of “cybernetics” in the 1960s-70s – but with a background in anthropology and psychology rather than computer science – Bateson has some prescient perspectives on AI and its relation to the concept of “mind” in general, such as the following (apologies for the gendered language):
“Now, let us consider for a moment the question of whether a computer thinks. I would state that it does not. What ‘thinks’ and engages in ‘trial and error’ is the man plus the computer plus the environment. And the lines between man, computer, and environment are purely artificial, fictitious lines. They are lines across the pathways along which information or difference is transmitted. They are not boundaries of the thinking system. What thinks is the total system which engages in trial and error, which is man plus environment.”
Let’s stop and reflect on that, because it’s quite a challenge to conventional ways of thinking. In Bateson’s model, a mind is made up of those “pathways along which information or difference is transmitted”. He specifies that mental process requires “circular or more complex chains of causation”, i.e. feedback loops which allow for iterative learning; and he highlights that the mind, the “total self-corrective unit which processes information”, is a “network [that] is not bounded by the skin but includes all external pathways along which information can travel”1.
In other words, our own minds (as human individuals2) are not only inside our brains – or even just within our bodies – but encompass all of the messages that we receive and those we transmit through our actions and communications, and the cyclical processes of causality through which those signals interact with the world around us, including both the “natural” and “artificial” aspects of that world.
But in the title of this post, I have drawn a distinction between artificial and natural intelligence. Is that a false distinction? Are we all – humans, other living organisms, environments, computers and whatever else – simply overlapping sub-units of one massive mind?
Map, meet Territory
As far as I can see, we are all indeed participating together in a vast mental process – whose “pathways” transgress all kinds of commonly imagined boundaries (living / non-living, natural / artificial, me / you, etc) – and yet, I believe there is an important distinction to be made between the “intelligence” of AI and that which we experience at first hand. It’s one that Bateson made too, in his oft-quoted words (originating with Alfred Korzybski), “The map is not the territory”.
The difference – I think – between the intelligence of an AI system and that of a human or other living being is that AI can only handle the map, in other words the model or representation of a thing or situation. It has no access to the territory, the reality, the presence itself.
I’m reminded of a recent Note by Henrik Karlsson:
“When drawing, the trick is to spend more time looking at the thing you are trying to draw than at your picture. This is true for writing, too.”
AI can only look at the “picture”: the data it has been given. It has no way to perceive directly the “thing it is trying to draw”. No wonder it’s vulnerable to hallucinating!
Of course, humans are susceptible to such problems too – and increasingly so, the more that our civilised, tech-enabled way of life separates us from the world of real bodily experience, and replaces it with simulacra designed to reinforce existing preferences (feeding addictive habits of content-consumption; keeping us in “bubbles” or “echo-chambers” of those who agree with us), which often work to exacerbate the most established prejudices:
“Generative AI is structurally conservative, even nostalgic”, Prof Roland Meyer comments in this article, which goes on to explain: “generative AI is trained on pre-existing data, which research has shown is inherently biased against ethnic diversity, progressive gender roles and sexual orientations, therefore concentrating those norms in the output”.3
What alternative is there? As humans, are we actually capable of perceiving “the territory”, or are we, like AI, confined to living in and reproducing our maps?
Bateson offers a skeptical take on our ability to know reality directly, pointing out that even our sensory impressions (sight, for example) are maps of a kind:
“We say that the map is different from the territory. But what is the territory? Operationally [in making a literal map], somebody went out with a retina or a measuring stick and made representations which were then put upon paper. What is on the paper map is a representation of what was in the retinal representation of the man who made the map; and as you push the question back, what you find is an infinite regress, and infinite series of maps. The territory never gets in at all.”4
However, he goes on to attempt to delineate what he calls “the other way of thinking”, stating that taking LSD had helped him to imagine what a different habit of thought might be like: “Under LSD, I have experienced… the disappearance of the division between self and the music to which I was listening. The perceiver and the thing perceived become strangely united into a single entity. This state is surely more correct than the state in which it seems that ‘I hear the music’.”
This is the direction I was pointing towards when I contrasted representation with presence itself. Here I am influenced by Iain McGilchrist’s use of the verb “presencing” to describe what the right hemisphere of the brain does and “re-presenting” as the work of the left hemisphere, a distinction summarised in these paragraphs:
“I suggested that there were two ways of being in the world, both of which were essential. One was to allow things to be present to us in all their embodied particularity, with all their changeability and impermanence, and their interconnectedness, as part of a whole which is forever in flux. In this world we, too, feel connected with what we experience, part of that whole, not confined in subjective isolation from a world that is viewed as objective. The other was to step outside the flow of experience and ‘experience’ our experience in a special way: to re-present the world in a form that is less truthful, but apparently clearer, and therefore cast in a form which is more useful for manipulation of the world and one another. This world is explicit, abstracted, compartmentalised, fragmented, static (though its ‘bits’ can be re-set in motion, like a machine), essentially lifeless. From this world we are detached, but in relation to it we are powerful.”5
The Master and his Emissary was first published in 2009 – long before AI entered the mainstream – but these descriptions of how the brain functions when the left hemisphere takes the lead (stepping “outside the flow of experience… to re-present the world”) seem to me to be a fitting description of how AI works too.
Back to the present?
Despite – or perhaps because of – the immense power and resources invested in artificial intelligence (both digital AI and the older version embodied in our left-brain-led forms of communication and social organisation), and the increasingly evident forms of distortion, error and untruth it can manifest, there are signs of a counter-movement: of people finding ways to re-establish our ability to “presence”, to become aware of the “embodied particularity” of things and of the “whole which is forever in flux”.
The resurgence of psychedelics is one path in this direction (which I mention here because of the influence of LSD on Bateson’s thinking – though it’s not a route I would choose myself, having experienced psychedelic-adjacent states spontaneously).
The various practices of “embodiment” and “somatics” – ranging from dance and vocal work to therapeutic modalities – are another way to point our awareness away from what the (left) brain thinks it knows towards the aspects of our being and experience that remain unlabelled and continuously emergent. Similarly, I’ve always had an interest in craft practices and the experience of making by hand, with the relationality between human and materials this entails.
Enquiries that come under the heading of “animism” are also of relevance here, and are a specific focus for me this year, as I participate in
and ’s Whole Earth Animism course, and am currently reading Robert Macfarlane’s new book Is a River Alive? He quotes Ursula Le Guin as saying: “One way to stop seeing trees or rivers or hills only as ‘natural resource’ is to class them as fellow beings - kinfolk. I guess I’m trying to subjectify the universe, because look where objectifying it has gotten us”.Here on Substack, an essay I particularly enjoyed on this topic was this from Erik Jampa Andersson, who suggests that “A tree spirit is perhaps, quite simply, the subjective ‘person’ embodied in a tree. They are not a ‘person’ in precisely the same way that we humans are ‘persons’, but they are persons nonetheless — complete with their own sensory experiences, memories, desires, and inner lives. They are, in a sense, ‘inspirited’.”
With, not about
My recent Ecologia post showed the results of some explorations into thinking with (rather than thinking about) the world around me – specifically, the weather.
Re-engaging with Bateson is helping me see more clearly the significance and urgency of such a shift in perspective. He frames it as a matter of evolutionary survival:
“In accordance with the general climate of thinking in mid-nineteenth-century England, Darwin proposed a theory of natural selection and evolution in which the unit of survival was either the family line or the species or subspecies or something of the sort. But today it is quite obvious that this is not the unit of survival in the real biological world. The unit of survival is organism plus environment. We are learning by bitter experience that the organism which destroys its environment destroys itself.”
This calls to mind a song I learned back in February, on a singing retreat with Clara Clay:
“We are the river, and the river is us.
We have no choice but to stand up.
And the river holds our story,
And the story holds our people,
And the people hold the river of my life”6
An almost identical statement is quoted in the Te Award Tupua Act passed by the parliament of Aotearoa New Zealand in 2017, which recognises the “legal personality” of the Whanganui River. The Whanganui people have a proverb (referenced by Macfarlane7): “Ko au te Awa; ko the Awa ko au”, translated as “I am the river; the river is me.”
Bateson warns that the contrary perception – of ourselves as separate from the river, or the wider ecosystem we depend on – is not only mistaken, but dangerous:
“… When you separate mind from the structure in which it is immanent, such as human relationship, the human society, or the ecosystem, you thereby embark, I believe, on fundamental error, which in the end will surely hurt you.”
He proposes that “the most important task today is, perhaps, to learn to think in the new way” - but admits that “I don’t know how to think that way… if I am cutting down a tree, I still think… I am cutting down the tree. ‘Myself’ is to me still an excessively concrete object, different from the rest of what I have been calling ‘mind’.”
Making a different way of thinking – an integration of mind with world – habitual is not easy, while so much of our culture relies upon and reinforces the belief that we are separate and in competition.
Among the tasks of my writing on Substack is to help me imagine what a different, more integrated way of thinking might be like – and begin to establish it as a habit, and as something shareable. You may have noticed that this publication’s title, At the Water’s Edge, is a poetic image of this threshold: the “Water” evoking the ever-flowing reality of life in all its presence (as the right brain knows it); the “Edge” suggesting where it meets the seemingly more solid (yet actually more illusory!) world of conceptual structures, familiar and graspable to the left brain.
If you are in a similar process of learning to perceive differently and to act in accordance with our interdependence, I’d love your company here. Please like / share if you found this article helpful, and drop a comment below to let me know what resonated in particular.
Steps to an Ecology of Mind, p.319
Bateson acknowledges the validity of the concept of an individual mind, distinct from the wider world, but suggest that its definition and boundaries must be contextual: “The delimitation of an individual mind must always depend upon what phenomena we wish to understand or explain. Obviously there are lots of message pathways outside the skin, and these and the messages which they carry must be included as part of the mental system whenever they are relevant” (my emphasis).
For a challenge to such norms check out
, who claims to have developed “the world’s first Decolonial, ethical, culturally competent AI Framework that identifies all bias, and deconstructs it”.Steps to an Ecology of Mind, p. 460
The Master and his Emissary, p.93
Based on the words of LaDonna Brave Bull Allard, Founder of Sacred Stone Camp at Standing Rock, as quoted in this article.
Is a River Alive? p.28