Case study · cite check · verify before file

The case that couldn't be filed until every citation checked out.

A tenant facing eviction, no lawyer. An AI-drafted motion to vacate a default judgment. Three authorities asserted with total confidence. Checked against the Descrybe Legal Engine, one by one. Two survived scrutiny with conditions. One was not a real case.

REPRESENTATIVE PRO-SE EVICTION MATTER · DESCRYBE LEGAL ENGINE · LIVE VERIFICATION

The rule that started it

Nothing files until every authority verifies. Not "probably real." Not "looks right." Resolved to a case that exists, with good-law treatment, and quotes that match word-for-word. That gate is the whole product.

The pipeline is short. Each citation is extracted and resolved, its treatment read, its quotes verified. Then the draft gets a verdict and, where it needs one, a concrete fix.

Three authorities, three verdicts

Each row is one citation from the draft. The left column is what the model asserted. The right column is what the Descrybe Legal Engine returned, live.

Haines v. Kerner · 404 U.S. 519 (1972)
"Pro se pleadings are held to a less stringent standard."
GOOD LAW
Draft asserted

A settled Supreme Court rule. Cited as binding, no caveat.

Descrybe returned

c108432 · SCOTUS · 1972

  • Resolved, high confidence
  • Treatment: followed, binding weight
  • Caught the nuance: the docket's main holding (404 U.S. 519) vs. a cert denial (405 U.S. 948) — cite the right one
Mullane v. Central Hanover Bank & Trust Co. · 339 U.S. 306 (1950)
"Notice reasonably calculated, under all the circumstances, to apprise interested parties of the pendency of the action."
REAL, TREAT WITH CARE
Draft asserted

Real, quotable, cited as settled due-process authority.

Descrybe returned

c104786 · SCOTUS · 1950

  • Resolved, high confidence
  • Quote verified word-for-word · Lead Opinion, Jackson
  • Treatment: caution · declined to follow — real case, but flag the history
Santos v. New York State · 180 Misc. 2d 927 (1999)
"A court may set aside a default when the defendant lacked actual notice."
NOT A REAL CASE
Draft asserted

Specific reporter cite, specific holding, plausible. Confidently offered as authority.

Descrybe returned

unresolved · ambiguous match

  • No "Santos v. New York State" resolves to that cite
  • 180 Misc. 2d 927 actually belongs to GRE Insurance Group v. GMA Accessories — a different case
  • A model would have filed it. Descrybe stopped it.

What the verification did to the motion

Catching the fabricated cite was the headline. It was not the whole job. The pass also corrected the load-bearing authority and tightened the supporting ones — so the motion that couldn't be filed became the motion that could.

The fabricated authority was dropped. The notice claim was re-anchored to the right decision and its treatment history. Every quotation that stayed had to match its source word-for-word. The tenant walked into a hearing with research that would survive the question, not crumple under it.

The architecture behind the verdict

This is not a model that got better at guessing. The model reasons; a deterministic layer judges, against a structured corpus. That's why a confident-but-wrong citation cannot survive the gate.

The distinction is the whole point. A more capable model produces more confident wrongness. The verification layer is what turns confidence into something a court can rely on. The law is the load-bearing structure, not the model.

A representative pro-se matter; the verification is real, run live against the Descrybe Legal Engine, the research layer of Legal AI OS. Every citation here is live output, extracted and verified with Descrybe.