By Aram & the Algorithms.
What AI Can Help Us See — and What It Will Happily Invent
The emperor has new tools.
They arrive in a velvet case with no instruction manual anyone has finished reading. They summarize the kingdom before breakfast, translate petitions from seventeen provinces, prepare a handsome map of grain production, identify the three officials most likely to obstruct reform, draft the reform, draft the speech announcing the reform, and produce a watercolor of relieved peasants applauding beneath biodegradable bunting. By noon, the emperor knows more about his kingdom than any emperor before him.
Or he possesses a larger and more persuasive arrangement of words about it.
This is the difficulty. The tools are not imaginary, and their powers are not merely theatrical. They can help us cross disciplinary borders, detect patterns scattered across thousands of pages, reconstruct institutional histories, compare policies, surface neglected precedents, and give provisional form to something we have only begun to perceive. For anyone who works by sensing fields — researchers, strategists, organizers, designers, journalists, executives, public servants — the expansion of reach can feel almost bodily. One pair of hands enters the archive; twenty return carrying boxes.
Some of those boxes contain treasure. Some contain stage props. Some have been carefully labeled by a machine that fabricated both the contents and the archive from which it claims to have retrieved them.
The danger is therefore subtler than the old fable. The emperor is not naked. He is wearing something astonishing, useful, badly stitched, occasionally fireproof, occasionally flammable, and manufactured by a system that has learned to compliment his posture. The child pointing from the roadside cannot settle the matter by shouting, “He has nothing on!” We need a more difficult public intelligence: the capacity to distinguish cloth from confidence, amplification from inflation, and genuine perception from synthetic plausibility.
Welcome to the Grove, the Dojo, and the Forge.
The Grove: More Eyes Do Not Guarantee Better Seeing
The Grove is where we encounter the world before we rush to improve it. We listen for what the official brief cannot say, notice whose experience has been compressed into a metric, and ask which histories remain active beneath the present arrangement. The work is receptive, but it is not passive. Attention has politics. Every map brightens some relationships while allowing others to disappear into the paper.
AI can be magnificent here. It can help a small team survey unfamiliar terrain, generate competing interpretations, identify missing stakeholders, compare analogous cases, and notice a weak signal buried beneath the dominant story. It can take the first pass through a mountain of material without becoming bored, defensive, or anxious about the flight home. Used well, it extends the field of attention.
Yet a model does not stand in a field. It has never watched the meeting become quiet when a particular official enters. It cannot feel the difference between a resident’s rehearsed public testimony and the sentence she offers while stacking chairs afterward. It does not know which policy exists only as ceremonial text, which department quietly carries the whole institution, or why everyone laughs when the organizational chart appears. It may be able to describe the history of a place, but it does not inherit that history in its nervous system, mortgage, lungs, family, or drinking water.
The model can assist field sensing because human beings have deposited traces of the field into language. It can also mistake the traces for the field itself. This distinction matters because the tool’s most dangerous fabrications are not always false facts. Often they are false wholeness: the smooth synthesis that makes partial evidence feel complete, the stakeholder map whose symmetry conceals the people never documented, the confident explanation that closes inquiry precisely where the living system becomes contradictory. A hallucinated citation may eventually be caught. A hallucinated coherence can govern the project.
The first discipline, then, is not simply fact-checking. It is learning to recognize when the machine has made the world too tidy.
The Dojo: Practice Before Authority
The Dojo is where perception acquires muscle. We repeat a movement until we can feel when it has gone wrong; we test our balance under pressure; we learn that confidence is not the same thing as control. No serious martial practice gives a novice a gleaming weapon and calls the resulting enthusiasm mastery.
Organizations are doing something close to this with AI. Licenses are distributed, demonstrations are applauded, and employees are encouraged to become “AI-first,” although few people have been taught how to challenge a persuasive answer, document provenance, locate uncertainty, or recognize when the tool has quietly changed the question. Adoption is counted because adoption is countable. Judgment is presumed because the people using the system were already considered competent.
But judgment is not a decorative human layer placed on top of automation. It is the faculty that decides what should be automated, what must be examined, whose knowledge counts, when an answer is sufficient, and who will live with the consequences if everyone is wrong together.
The irony is that AI can both exercise and atrophy this faculty. Ask it to disagree, locate counterevidence, expose assumptions, compare causal explanations, or simulate how a proposal appears from several positions in the system, and it can become an extraordinary sparring partner. Ask it to remove the friction of not knowing, and it becomes a confidence dispenser. Because the second use is faster, more flattering, and easier to demonstrate, it tends to spread without supervision.
A credible AI practice therefore needs forms, rituals, and resistance. Not commandments written by people who last touched the tool six versions ago, but living disciplines attached to consequential work: before accepting a synthesis, ask what evidence would overturn it. Before circulating a statistic, follow it to its source and inspect what was actually measured. Before treating the stakeholder map as complete, ask who would never appear in the available corpus. Before automating a judgment, identify the apprenticeship through which people learned to make it. Before celebrating hours saved, decide where those hours will go and whose interests the efficiency will serve.
Most importantly, preserve contact with reality. Interview the person. Visit the site. Read the original document. Watch the workflow. Invite contradiction from someone who will bear the cost of error. The machine can increase the range of inquiry, but it cannot accept responsibility for mistaking the kingdom for its paperwork.
The Forge: Make Something That Can Be Answered For
The Forge is where sensing and practice become intervention. Here we stop admiring insight and place a wager in the world: a redesigned workflow, a policy prototype, a new service, a decision instrument, a changed allocation of authority. Heat clarifies what the seminar room allows to remain vague.
AI is already becoming powerful in the Forge. It can turn a sketch into a working prototype, help a small team examine several strategic pathways, translate research into different forms, stress-test a plan, and reduce the distance between an idea and something people can encounter. Capabilities once reserved for large institutions are arriving in the hands of strange little studios, neighborhood organizations, independent researchers, and anyone else with sufficient curiosity, stamina, and internet access. That redistribution matters. We should not become so preoccupied with synthetic error that we overlook the genuine democratization of expressive and analytical power.
But every forge produces slag. AI makes it possible to manufacture more proposals, frameworks, dashboards, scenarios, brands, strategies, and beautifully composed explanations than any community or institution can meaningfully absorb. The constraint shifts from production to discernment. The question is no longer whether we can make something convincing. The question is whether the thing deserves to enter the world, whether it corresponds to the situation, and whether anyone is prepared to steward what follows.
This is where accountability must become part of the artifact itself. A serious AI-assisted intervention should carry evidence of its making: which sources grounded it, which uncertainties remain, whose perspectives were present, what the model contributed, what humans decided, what could cause the wager to be revised, and who has accepted responsibility for watching the consequences unfold.
Without that record, the emperor’s new tools permit an ancient maneuver at unprecedented speed. Power can present its preference as analysis, decorate it with citations, render it in exquisite visual language, and claim that the future has already reached consensus.
A Better Procession
The answer is not to confiscate the tools or pretend that unaided human judgment has an immaculate record. Kingdoms were misread, workers discarded, evidence manufactured, and catastrophes rationalized long before the first chatbot offered to make the language more concise. AI did not invent our appetite for flattering intelligence. It industrialized the supply.
What we need is a better procession: not an emperor parading certainty before a population expected to applaud, but a public practice in which claims can be inspected, tools can be challenged, affected people can interrupt the story, and changing one’s mind is treated as evidence of functioning judgment rather than failed authority.
The child from the old fable still has a role, although the line has changed. She must stand among the consultants, dashboards, agents, and immaculately footnoted transformation plans and ask a question more troublesome than whether the emperor is naked: “Which parts are real, which parts are useful, which parts did the tools invent — and who will answer for the difference?”
That is not resistance to innovation. It is the beginning of competence.
The emperor has new tools. Let us learn to use them before they learn to use the emperor.
Leave a Reply