Hallucinated certainty
The presentation of a genuinely contested legal question as settled, such as a circuit split rendered as a single rule, in output whose citations may all be real and accurately quoted.
First used in The Four Laws of System Design for Computational Law
Hallucinated certainty is the Institute's name for the failure its gray-area papers call false confidence, and the subject of its first thesis: hallucinated certainty is more dangerous than hallucinated citations.
A fabricated case announces itself; opposing counsel cannot find it and a citator does not recognize it. A contested question presented as settled announces nothing. A system can retrieve accurate authority on both sides of a split and still synthesize a confident statement that the law is one side. It has not hallucinated a case. It has failed to surface the ambiguity.
The Zeroth of the Four Laws of System Design targets the failure directly: a system must never present outputs with unwarranted confidence. The remedy is detection rather than blanket hedging: telling a gap in the research apart from a gap in the law, and saying specifically where courts divide.
Related terms
Related frameworks
The Four Laws of System Design for Computational Law
A legal AI system must never overstate its confidence, must explain every output, must ground every legal proposition in validated authority, and must keep its reasoning components testable and versioned, with each law yielding to the laws above it.
Grayness Score and Gray Area Radar
The Grayness Score is a composite, evidence-linked indicator of how far the legal system itself treats a question as contested, and the Gray Area Radar is the view that shows a lawyer which signals fired and where they came from.