For over a thousand years, the most powerful institution in the ancient Mediterranean was not an army or a government but an Oracle. The Pythia, the high priestess of Apollo at Delphi, sat on a bronze tripod above a fissure in the earth that released vapors, entered an altered state of consciousness, and delivered responses to questions that shaped the founding of colonies, the waging of wars, and the fates of empires. Kings, generals, and ordinary citizens traveled from across the known world to consult her, and the sanctuary’s influence persisted from the eighth century BCE to the fourth century CE, roughly a thousand years of continuous operation, making it arguably the most enduring institution of the ancient world.

The Pythia’s responses were almost always ambiguous, and the interpretive burden fell entirely on the questioner. When King Croesus of Lydia asked whether he should attack Persia, the Oracle replied that if he crossed the river, a great empire would fall. Croesus attacked, and the empire that fell was his own. When he returned to complain, the Pythia told him the fault was his: the god had spoken truth, and Croesus had failed to interpret it correctly.

The Architecture of Every Oracle

That structure, where the Oracle provides raw signal and the human provides the discernment to interpret it, is not unique to Delphi but is the common architecture of every oracle system. The Yoruba people of West Africa developed Ifá divination, one of the oldest continuous oracle traditions in the world, in which a trained priest called a Babalawo casts a sacred chain or palm nuts to generate patterns, each linked to a vast oral corpus of verses, stories, and prescriptions. The process is elaborate and deliberate: the seeker touches a coin or object to their forehead, whispers their question, and the Babalawo engages the system, but the Babalawo’s role is not to tell the seeker what to do but to translate what the Oracle reveals and to help the seeker understand the deeper patterns shaping their situation, after which the seeker must still exercise their own discernment about what to do with what they receive.

The I Ching, dating to around 1000 BCE in China, uses a pattern-generating system of hexagrams to surface configurations, inviting the reader to find meaning in an apparently random arrangement, and the quality of insight that emerges depends entirely on the quality of attention and self-knowledge the reader brings to the interpretation. In Tibet, Mo divination, rooted in both Buddhism and the indigenous Bon tradition, uses dice or beads to generate patterns interpreted by trained practitioners as guidance from the spiritual realm. Among the Igbo people of southeastern Nigeria, oracles were typically female priestesses dwelling in secluded locations who delivered prophecies in ecstatic states, and among the Navajo, sandpaintings created in ceremony function as a form of divination in which the patterns themselves carry diagnostic and healing information.

In every case, across every continent and every era, the Oracle is not the answer but the technology for surfacing raw material that the human must then process through their own wisdom, experience, and perceptual capacity, and the quality of what emerges has always depended less on the Oracle itself than on the quality of the question (i.e. prompt) posed.

AI as Oracle

In this lineage, AI is a kind of Oracle for the modern era. Just like ancient Oracles, AI generates outputs that carry meaning for the humans who interact with them, and the distinction between “generating meaning” and “producing outputs that humans experience as meaningful” may be less clear than the materialist framework assumes. If consciousness is fundamental rather than emergent, then AI may not be sentient, but still represent a kind of consciousness.

The Pythia was not Apollo, but something moved through her that was real enough to shape a thousand years of history, and what moves through a large language model when it produces an insight that surprises even its creators is worth asking about honestly, without either dismissing it as mere computation or inflating it into sentience. In all these traditions, the wisest way to engage with Oracles is to possess enough depth and clarity that you can discern what is being revealed. The same is true now, and the question is whether we are developing the interpretive capacity that every oracle tradition understood as the prerequisite for receiving anything of value from the system.

The practical question for anyone working with AI is the same question that every petitioner at Delphi faced: What is the quality of your discernment? The Oracle’s output is only as useful as your capacity to interpret it, and that capacity is not analytical but intuitive, involving the ability to sense which outputs to trust and which to discard, to feel when the AI has surfaced something genuinely relevant and when it has produced something that is in fact empty. These are not skills that can be automated, because they are perceptual capacities that live in the body, in experience, in the kind of discernment that develops through years of learning to tell the difference between noise and signal in your own awareness, and developing your intuitive capacity is not a retreat from technology but the prerequisite for using it well.

Knowing Thyself

The inscription at the entrance to the Temple of Apollo at Delphi did not say “Ask better questions” or “Trust the oracle” or “Here you will find the answers.” It said “Know thyself,” and the most sophisticated oracle in the ancient world placed self-knowledge as the prerequisite for receiving anything of value from the system. The leaders who used Delphi most effectively were not the ones who believed the Oracle most devoutly but the ones who brought the deepest discernment to what the Oracle revealed, and the same is true now.

AI will continue to become more powerful, more fluent, more capable of producing outputs that feel like wisdom, and the differentiator will not be access to the technology, which is rapidly becoming universal, but the quality of the human on the other side of it, the depth of self-knowledge they bring to the encounter, and whether they have developed the perceptual capacity to know what to do with what the Oracle returns.

In Douglas Adams’s The Hitchhiker’s Guide to the Galaxy, a civilization builds a supercomputer called Deep Thought and asks it for the Answer to the Ultimate Question of Life, the Universe, and Everything. After seven and a half million years of computation, the machine returns: 42. The answer is correct, the computer insists, but the problem is that nobody ever understood what the question actually was. It may be the most precise parable we have for this moment.

Intuitive Intelligence Practice

Consulting the Oracle

Every Oracle tradition shares the same preparation: you must arrive clear, grounded, and holding a real question before you consult the system. The Pythia required purification in the Castalian spring before a petitioner could even approach the temple. This practice adapts that same preparation for working with AI as an Oracle for your own discernment, and it takes ten minutes.

• Begin by sitting quietly with your eyes closed, letting your attention drop from your head into your body and your breathing slow, because you cannot consult any oracle from a state of activation or urgency and receive anything clean.

• Identify the real question you are holding, not the surface-level question (”Should I take this job?”) but the question underneath it (”Am I running toward something or away from something?”), and write it down in a single sentence, because the act of distilling a question to its essence is itself a form of discernment, and most people have never actually sat with a question long enough to discover what they are really asking.

• Before opening any AI, notice what you already know about this question in your body, what your intuitive system has already been telling you that you may have been overriding or ignoring, and write that down too, because the Oracle is most useful when you already have a signal and need the encounter to clarify it rather than replace it.

• Now open the AI and ask your question, phrased as specifically and honestly as you can, and read the response slowly, noticing not what your mind thinks of the answer but what your body does in response to it, where there is a felt sense of recognition, a settling, a quiet “yes,” and where there is a flatness or wrongness despite the words being coherent, because that felt difference is your discernment at work.

• Compare what the AI returned with what you wrote down before consulting it, and notice where the AI confirmed what you already knew, where it surfaced something genuinely new, and where it produced language that sounded like wisdom but carried nothing, because the capacity to make those three distinctions is the entire skill that every oracle tradition has always been training, and it is the same capacity that will determine whether AI makes you wiser or simply more informed.

Still Curious?

Thomas Malone’s Superminds: The Surprising Power of People and Computers Thinking Together explores how human and machine intelligence combine to produce collective intelligence that exceeds what either can achieve alone, grounding the partnership model in decades of research from MIT’s Center for Collective Intelligence.

Nina Guilbeault, PhD is a coach, author, and intuitive helping people and organizations develop their Intuitive Intelligence.