Directional opacity
The inability of a legal research system to represent or query which party a proposition favors, and at which procedural stage, so that every result arrives as neutral topical content the attorney must sort by hand.
First used in The Promise Fulfilled
Directional opacity is one of five structural failure categories the Institute identifies in document-level legal AI, and the most direct expression of a long-standing gap. The AI and Law tradition treated direction as foundational: HYPO modeled factors that strengthen or weaken a party's position, and CATO tagged each factor as plaintiff-favoring or defendant-favoring.
A system without a directional field cannot answer a query such as "defendant-favorable authority on foreseeability at summary judgment in California premises liability." Direction is also subtler than who won: a case can announce a rule that favors defendants while the plaintiff prevails on another ground.
Unlike some failures, this one is not mitigated by agentic web search, because no retrieved document contains the missing field. It has to be classified at ingestion.