About This Article
This is Article 1 in The Missing Normative Layer series. It introduces the structural problem at the base of AI adjudication: ethics and normative values cannot support coherent adjudication until their underlying elements have been properly identified and made usable for reasoning.
The Structural Prerequisites of AI Adjudication
AI is facing a problem that cannot be solved merely by placing limits around behaviour. In adjudicative contexts, ethics and normative values cannot simply be added through good intentions, hard-coded constraints, or static rule sets. They must first be decomposed into formally usable building blocks.
AI adjudication begins earlier than most discussions assume. Before a system can reason over ethics or normative values, the normative field itself must be rendered into a form that reasoning can actually operate on.
1. Identifying the Building Blocks
The first requirement is definitional. Before a system can reason over ethics or normative values, the normative field must be decomposed into elementary units that can be identified with precision. Broad value language is not enough. Terms such as fairness, dignity, responsibility, or harm often appear as single concepts, but in practice they contain multiple underlying components, including obligations, prohibitions, permissions, priorities, conditions, and thresholds. As long as those components remain bundled inside vague, composite, or rhetorically framed expressions, the normative field itself remains under-defined. What appears to be clarity at the surface often conceals unresolved structure underneath.
Until normative values have been decomposed into atomic form and given an explicit relational structure, coherent ordering of normative priority is not possible.
2. Formalizing the Building Blocks
The second requirement is formalization. Once the normative building blocks have been defined in sufficiently elementary form, they must be expressed in a way that allows logical operators to act on them as they would on any other element within a reasoning structure. Only then can they be reasoned over and operate naturally within complex chains of reasoning. This is what makes it possible for normative elements to interact with competing constraints, participate in adjudicative logic, and contribute to traceable outcomes without reliance on hard-coded rules.
That is where the real challenge lies: it is not merely to make AI behave acceptably within a known normative field. It is to make the field itself structurally usable for reasoning. Until that becomes possible, coherent AI adjudication cannot be realized.
A Practical Domain of Normative Conflict
Artificial gestation is one example of the kind of adjudicative domain this problem points toward. It brings multiple normative layers into direct conflict, including questions of personhood, parental status, bodily autonomy, developmental interests, medical intervention, and institutional responsibility. It therefore illustrates why normative values cannot remain vague, bundled, or informally handled if they are to support coherent adjudicative reasoning.
What makes such cases especially significant is not only their ethical sensitivity, but their structural density. Multiple claims, duties, thresholds, and competing interests converge at once, often without any settled ordering between them. That is precisely the kind of condition in which informal judgment begins to break down and the need for explicit normative structure becomes impossible to ignore.
This is Article 1 in The Missing Normative Layer series. Article 2 turns to artificial gestation as a practical case and examines the normative conflicts it concentrates within a single adjudicative domain.
