AI Assistance

On help at hand.

AI systems that can read, organize, and cross-check dense, symbol-laden, self-referential text are already doing intrinsic, concrete work on this material. What follows is an account of that assistance, kept modest on purpose. Overstating it here would undercut the discipline EIT exists to bring.

• AI ASSISTANCE: The use of AI systems to support work with Conciliatorics. What that support can do is graded by readiness. The interpretive step remains Contemplation, and remains with the intervener.

Schematic Representation is the natural interface. Discursive prose has to be parsed for meaning, symbolic notation interpreted against a key, a diagram visually recognized. A schematic record, built from explicit fields and stated relationships, can be read and checked the way a table is read rather than a paragraph decoded. That is why AI assistance and Schematic Representation belong in the same jurisdiction: the representation type an AI can hold most accurately is also the one best suited to being acted on reliably.

AI Assistance Readiness

Assistance with this material can be graded by how far a use has been shown to work: validated, derived, or theoretical. A theoretical use should not be read as a proved one.

Stage A — Validated. Uses already proved useful and correct. As terms, symbols, and jurisdictions accumulate, an AI can check that a new symbol does not collide with an existing one, that a term’s use matches its definition, and that cross-references stay accurate — exhaustive work for a human to sustain by hand, and straightforward for an AI without fatigue. This has already happened. The Synthesis symbols originally read oS / iS / mS / uS / cS, and cS collided with an unrelated symbol from Inventive Edification. Once every symbol sat in one schematic glossary, the clash was visible, and the whole set moved to Σ. An AI can now run that class of check without fatigue.

Stage B — Derived. Uses that can already be put into practice: feasible, not yet done. They follow from the validated work. If an AI can check a glossary record against a schema, it can check a case record the same way. When a real case is being analyzed — a policy, an organization, a historical episode — the reading can be held as a schematic record: which jurisdiction, which pairs, which positions on them. A useful discipline, still informal, is treating certain pairs as mandatory: if a case is read through Fronterization and Permeability, a record with only one side filled in should read as incomplete. Enforcing that completeness is the kind of check a schema — and an AI checking against it — can hold a human analyst to. A case library built this way has been named. It has not been built.

Stage C — Theoretical. Speculative uses that may be helpful, and are not tested as feasible yet. A small derivation engine — a program that takes a completed schematic reading and mechanically checks or derives what the Attributes Status equations say should follow from it — may eventually be possible. It would still only run on judgement a human had already supplied. It would not perform the categorization. No such engine exists yet, and nothing here claims that one is imminent. It is on the record as a possible direction, alongside whatever other applications its readers bring to the parcel.

AI Contemplation

None of the three stages is the interpretive step. Deciding that a given policy is a matter of Permeability and Fronterization at all, and characterizing where it sits on each, is Contemplation: judgement applied to a specific, messy, real situation.

An AI can enforce that a reading is complete once it exists. It cannot supply the reading. Completeness belongs to the record. The reading belongs to the person who looked at the situation and chose. Stage A already needs a human to have introduced a term before a collision can be caught. Stage B needs a human to have chosen the pair before a missing side can be flagged. Stage C, if it ever exists, would still only run on a completed reading. Assistance starts after Contemplation has already been done.

That is not a gap to be engineered away, and it is not a limit that a later stage of readiness will lift. Contemplation is the inventive approach that emphasizes the intervener: knowledge-building and judgement. The intervener is the one who has to stand behind the reading — to revise it when a later case does not fit, to answer for it when someone else reads the same policy through a different pair. An AI system can hold the record of that work. It is not the intervener.

Transformation, which is EIT’s approach, emphasizes the system: concrete change and performance. Building a schema, checking a file, encoding a case: those are Transformation. They do not become Contemplation by being done well, or by being done by a more capable system. That limit follows from Contemplation as the approach concerned with the intervener’s knowledge-building and judgement — a role Transformation was never meant to take over.