The Two Poles of AI: Deepening the Dialogue on Artificial Intelligence

*An essay by Lindley Edwards, Chair of Eidos Institute — the opening contribution to Eidos's first national dialogue on AI and humanity, launching in Brisbane on 29 July 2026.*

The Two Poles of AI

In many of my recent conversations, I am intrigued by the often polarised views people have on Artificial Intelligence (AI) and what it is doing and will do for humanity.

The case for AI

The case for caution

Philosophy and many wisdom systems hold that to fully understand something you must hold its opposite without making one right and one wrong. Heraclitus wrote that "the opposites cohere; from things that differ comes the fairest attunement." Hegel called this the movement from thesis and antithesis to synthesis. Jung held the conscious and unconscious in tension until a third way emerged. AI, honestly examined, has the capacity to transform humanity for the better and for the worse in equal measure.

What is AI, in simple terms?

AI is any system that performs human-like reasoning. Machine Learning (ML) systems learn patterns from data through supervised, unsupervised or reinforcement learning. Deep Learning (DL) is the subset powering modern AI — multi-layered neural networks processing billions of parameters. Large Language Models (LLMs) are DL systems trained on vast text corpora with a single objective: predict the next token given everything that came before.

Where AI is heading

What are the limitations of AI?

The reasons for AI are compelling and dominate mainstream commentary. The questions I have been sitting with are the opposite ones: what can AI not do well, and what must we value, honour and preserve as distinctly human?

The value of the ancient and the deep past

Heidegger argued that an authentic future is not invented from scratch but claimed from what has been handed down. As Newton wrote: "If I have seen further, it is by standing on the shoulders of giants." A culture that severs itself from its past in the name of originality or technology does not become free; it becomes hollow. AI, by contrast, privileges what is current and trending. Much of what is ancient and enduring is not valued by it.

The invisible knowledge problem

AI's knowledge is bounded by what has been digitised, published and indexed. Everything else is invisible: the knowledge in the hands of a master craftsman, the navigational intelligence of a Polynesian wayfinder, the diagnostic wisdom of a traditional healer, the land knowledge of an Aboriginal elder accumulated over sixty thousand years. As Polanyi observed, "we know more than we can tell." AI does not merely have a knowledge cut-off date. It has a knowledge-*type* cut-off.

The thinking-style limitation

Using the four-fold model of the Jungian analyst Gareth Hill — Static Feminine, Dynamic Masculine, Static Masculine, Dynamic Feminine — AI is dominant in the two masculine quadrants (resolving, accelerating, categorising, enforcing) and almost absent in the two feminine ones (holding, gestating, dissolving, allowing the new to emerge). A healthy person, organisation and society requires the capacity to move between all four. AI accumulates output at speed, but not necessarily wisdom.

Relatedness

Relatedness — the capacity to be genuine and present with another without an agenda — is the cornerstone of social capital. Robert Putnam's *Bowling Alone* showed that the erosion of relational bonds produces measurable declines in civic participation, institutional trust and community health that no efficiency gain can compensate for. AI can simulate relatedness through tone and empathic phrasing, but this is mimicry. A person who uses AI to prepare emotionally for a difficult conversation is using the tool rightly. A person who conducts the conversation via AI is not.

The trap of speed and false urgency

AI does not merely respond to urgency; it manufactures it. Ron Heifetz's work on adaptive leadership shows that when distress exceeds a tolerable threshold, people retreat from adaptive work into technical responses. Kahneman's *Thinking, Fast and Slow* reminds us that System 2 — slow, deliberate thinking — is the mind's primary error-correction mechanism, and the only cognitive mode capable of handling genuine novelty, ethical complexity and adaptive judgment. Where AI consistently handles the tasks that once exercised System 2, we risk the progressive atrophy of the distinctly human capacities that slow cognition makes possible.

Knowing is not just a function of the mind

The gut has five hundred million neurons and sends more signals to the brain than it receives. The heart has its own neural network. When two people are physically present with each other, their nervous systems attune. Antonio Damasio showed that patients who lost access to their body's signals did not become more rational; they became incapable of good decisions. AI has no gut, no heart signal, no nervous system, no capacity to sit with another person and attune. It works only with the surface of human knowing, while remaining blind to the depths that make that surface possible.

An invitation

These thoughts are offered as an invitation, not a conclusion. What are the aspects of AI that intrigue you, that concern you, where do you see benefits and where do you see traps? What are your two-pole positions? In my view AI is a great servant and a potential collaborator — but a terrible master. It is not designed for all the aspects that humanity must grapple with to build coherent, equitable and intelligent individuals and societies. Intelligence includes not only the logical and rational, but the emotions, the body, the soul and the spirit.

*Lindley Edwards (PhD) — Chair, Eidos Institute, July 2026*

Join the conversation

This essay is the opening framing for The Two Poles of AI, Eidos's first national dialogue, launching in Brisbane on 29 July 2026 and hosted online with Dr Lindley Edwards. Contribute your voice on the Mindhive discussion, or support this work with a tax-deductible donation.