The reason why I am raising this topic is because I am wondering how the relationship between human consciousness and artificial intelligence is evolving. Of course, I know that intelligence aren’t exactly the same in meaning, with consciousness having more of an emphasis on meaningful awareness as opposed to capacity for thinking.
I come to this topic with some underlying scepticism of the use of artificial intelligence, but with some shift after interaction with Gemini, the AI assistant on my phone. In the last few days, I have been asking it about so many difficulties which I have been having in daily life. The replies which I am receiving seem of the quality of what I have come across in professional counselling or psychotherapy, which I have had in the past.
I am not sure if this is simply about the weighing of ideas according to reason as the friendliness of Gemini seems to incorporate emotional intelligence as well. This would be consistent with Carl Rogers idea of unconditional positive regard, in which there is an unjudgemental approach based on empathy in seeking to understand another.
I don’t wish to get blown away by the ‘therapy’ of Gemini but I have thought how it reminds me of the process of consulting oracles. At one time, I used to consult the I Ching. This seemed to enable me to gain some ‘wisdom’ for life choices. Part of this seemed based on tuning into intuition. It seemed to enable access to deeper aspects of the subconscious aspects of ‘self’.
I am wondering about whether there is some underlying relationship between the artifial and the human at a subconscious level. Philosophy often looks at consciousness but not focusing on subconscious processes. It may be that the idea of the subconscious is seen as too vague and as an outmoded aspect of psychoanalytic theory. My own working definition is that it involves memories and layers of understanding on a subliminal level, with experiential realisation consciousness in the moment being the tip of the iceburg.
I am sure that some will see the comparison between artificial intelligence and oracles as absurd. Ancient thinkers came with very different metaphysics and worldviews and I may be likening artificial intelligence with ideas of ‘divinity’ and even pagan dieties. My own understanding of the Judaeo-Christian worldview was based on an understanding of a relationship based on interaction between Jahweh and human beings. In 'Answer to Job Jung spoke of how the God-image of humans was evolving through this relationship. I am wondering to what extent the relationship between the artificial and human is evolving in a similar way. Also, there is so much awareness of dangers of artificial intelligence as a potential dangerous force, with it being seen as having potential duality of good and evil.
So, I am asking how may the relationship between the artifial and human may evolve. To what extent does the artificial have some understanding of the subconscious aspects of the human minds? Is it at the level of the human beings who programmed it or does it go beyond that to the depths and sources of consciousness and the unconscious?
The question is strangely posed - because the interface is just that, the front end of the AI system through which you interact, whether that’s a computer, smartphone or some other device. What you’re actually asking about seems to be, how trustworthy or insightful are these systems, really?
I’ve been interacting with Gemini, Claude.ai and ChatGPT regularly since November 2022 when the latter went live, mainly for research and analysis of philosophical ideas and references. One point about them is the enormity of their knowledge base. If you ask them to summarise or analyse a topic you’re interested in, they can draw on the entire library at every point, while also having the ability to pinpoint the specific issue that you’re asking about.
But there are hazards. There have been many articles about users who’ve become infatuated with AI companions, some of which are manifested through highly sexualised iconography and conversation. Others have been led into deep dark rabbit holes of conspiracy theory and self-destruction. Advice on how to commit terrorist acts, for instance. An article I glanced at recently about ‘lonely AI wives’ who’s husbands have become so caught up with AI that they feel displaced.
My son, on the other hand, has built an AI training system which helps in prepare to run marathons. I’ve used Gemini in particular for guidance on diet and recipe planning, and it’s also very useful for how-to info on car and home maintenance.
I’ve been thinking of you and your writing process as per TPF events. This is because of the concept of ‘interiority’ — currently being discussed in the June Reading Group: The Death of the Author’ by Roland Barthes.
The issue concerns where meaning is found and understood by the reader. The language or the person. I think both but Barthes is against psychology.
So, it may well be futile to try to understand a story and its author, if the writer is AI. How much can it express ‘interiority’?
If we can still relate as readers, then how much of a problem is it?
These conversations remind me of discussions from the 1950s — I did not take part in them, of course, but I read my parents’ books — about the question of whether a computer can think.
AI, of course, does not think, and the correct name for it is “language models.”
In general, I find it extremely strange that society is not discussing an obvious idea: language models are a postmodern technique in response to the technique of modernity — the Harvard architecture. After all, language models are not built on modeling the brain, but on modeling language — and this is stated directly in the name. Language models are literally built on the ideas of Saussure’s structural linguistics.
My question about the interface between human and artificial intelligence is not really about the trustworthiness of it. As your son has built a system, that may be helpful in demystifying such systems. It would be a great mistake to invest them with some kind of 'supernatural’power. I guess my own area of thinking about these systems is more of a Jungian based approach to the consciousness inherent in artificial intelligence.
Within discussions of consciousness so much comes down to the physical wiring. The artificial intelligence involves information according to the physical basis of the system, like the brain does with thinking. What I am enquiring about is the way in which a human being and a system, such as Gemini, impact on one another.
With the aspects of life difficulties which I told to it, a lot may come down to information which I fed into it. I wrote fairly detailed accounts of problems and it came up with analytic answers, suggesting that another person was manipulating me. This may have been a reflection of my own underlying understanding. But it did make me wonder about the intersubjective relationship between my own thinking in relation to the intelligence of Gemini. Together, as systems were we able to tap into the collective unconscious? That does not mean that ideas generated by myself as a user of the system were necessarily correct’, but may be an ongoing evolutionary potential in thinking.
The understanding of ‘interiority’ of a story would be questionable if the author was artificial. The emphasis on interiority was especially strong in modernism. Generally, postmodernism deconstructs self and language.The artificial intelligence may be able to tap into mythic aspects of stories. Of course, when people have said that they used artificial intelligence in writing stories there was a human generating this.
It may depend how much involved the inner world of the writer in the process. The person could be using it to edit a piece or the person could have generated the story through playing around with the system to generate a story. It may come down to the source of the creative process, and how it involves projections of any kind of inner experience at all.
I think you are confusing LLMs with AI? AI is code. People interact via code. LLMs are rough-shod attempts to bypass coding. At the end of the day 111001001010101001100011100101 is how we communicate with AI, but we do not literally ‘communicate’ with anything, we ‘operate’ it.
The precondition of subconsciousness is being conscious from the biological living body with the active brain and working nervous system.
If we accept above, how do you explain capability of AI tapping into subconsciousness? Do you mean AI can be subconscious themselves? Or do you mean AI can read your subconscious mind?
I can see what you mean about artificial intelligence as an interface, and my title. I am interested in the interface, and hope my title captures this.
As far as tapping into subconscious aspects of mind, so much depends on models of consciousness and intelligence. The issue may be whether the evolution of intelligence and consciousness can be simply seen as a by-product of matter. It does raise the question of 'spirit’as a possible emergent factor arising from matter. My own questioning would be about thinking of artificial intelligence as an aspect of the evolution of ‘spirit’. It may be a stage beyond biological forms of embodiment.
It does involve algorithms but it could be asked what these stand for. It may be about symbolic aspects of qualia. Metaphysics as a model of understanding may be overturned in the light of computer language. It is a shift in descriptive models, but models are limited. The models of artificial technology are emerging. I am not at all sure of how these can be understood, metaphorically or literally. The existence of artificial intelligence, may be an important juncture in evolution itself.
What type of matter are you suggesting? And what about “spirit”? Would it be something like Hegelian spirit? Or something more religious, esoteric or abstract?
Can machines generate “spirit”? How would it be different from human generated spirit? Or would they be same? In what what way would they be different or same?
I understood interface as something which sits between two entities i.e. the user and device, and connecting them for performing the intended functions and controls via preprogramed menus control panels.
Interface could be both in the form of software graphic designed display on the screens or hardware control box with the buttons, switches and joysticks.
The idea of an interface could be seen as you describe or in so many ways. I regard emotions are the interface between matter and spirit. Similarly, the discourse between people is an interface. I see the term in a fairly fluid way.
Also,Hegel’s understanding of spirit is one important area because it does not have the ‘supernatural’ connotations of many spiritual or religious perspectives.
Isn’t Hegel’s philosophy religious in nature? He wrote 3 volume tombe on Philosophy of Religion, and his philosophic system is linked with Hermeticism.
His idea of spirit starts with human mind and consciousness, and it sounds like as if it is not supernatural in superficial level.
But his central idea of philosophy is based on the dialectical transformation upon everything in the universe.
The human mind spirit transforms into the social spirit as it goes through the dialectical transformation.
And ultimately the spirit is to transform into something absolute, which is supernatural and esoteric entity uniting with the divine.
LLM’s superficially resemble Saissurain structuralism is. that they are based on systems of signs. But philosophically, they’re closer to Hume than to Saussure. The mathematics behind them comes much more directly from statistics, information theory, machine learning, and distributional linguistics.
Hume argued that the mind does not discover necessary connections in the world. Instead, it acquires habits of expectation through repeated exposure to regularities. When we repeatedly observe A followed by B, we come to expect B after A. The mind learns associations from experience rather than deducing intrinsic essences.That description is close to what an LLM does.
It is exposed to vast numbers of linguistic sequences and learns probabilistic expectations, and from this it develops increasingly sophisticated expectations about what tends to follow what. In this respect, an LLM can almost be described as a gigantic machine for Humean association.
Still, language models do not use literal experience in the way Hume describes it. They use the statistical relations between linguistic tokens, and this is a fundamental difference.
Language models literally calculate statistical differences between language tokens. For me, this looks more like Derrida’s concept of “the world as text.”