BriefCatch unveils RealityCheck to tackle AI errors

BriefCatch has launched a new verification product, RealityCheck, as courts face a growing number of legal filings with fabricated citations and other AI-generated errors.

RealityCheck flags hallucinated cases, fabricated quotations, and misstated holdings as lawyers draft and review documents. It also checks whether cited authorities actually support the legal propositions for which they are offered.

Judges in multiple jurisdictions have sanctioned lawyers after finding errors tied to the use of generative AI in briefing. BriefCatch points to a database maintained by legal researcher Damien Charlotin that has recorded more than 1,000 legal decisions involving AI hallucinations.

The solution is built into the drafting and review workflow and is also being made available to BriefCatch's federal and state court clients.

Court pressure

Courts have imposed monetary sanctions in cases involving incorrect authorities, and some decisions have led to lawyers being disqualified from matters. The launch comes amid broader debate in the legal profession about the safe use of AI tools in litigation and the reliability of citations and quotations produced by large language models.

BriefCatch positions RealityCheck as a way for firms and courts to catch errors earlier in the process. Ross Guberman, BriefCatch's founder and CEO, said court adoption would shift expectations for litigators.

"Once courts are running filed briefs through RealityCheck, the calculus changes for every litigator," said Guberman. "The question isn't whether to verify your citations. It's whether you want the court to find the errors before you do."

Two-step checks

RealityCheck uses a two-layer process. First, it validates citations to cases and federal statutes through database lookups, cross-checking reporter volume, court identifiers, and case names. BriefCatch said this step is designed to identify phantom citations and misidentified authorities.

Second, it uses AI models to assess quoted language and the connection between the cited authority and the proposition in the document. BriefCatch said the system checks whether a quotation appears in the underlying opinion and whether the authority supports the stated point. This approach is intended to identify fabricated quotations, misstated holdings, and unsupported propositions.

BriefCatch cited an example from a matter in the US Court of Appeals for the Fifth Circuit. Applying RealityCheck to a brief filed in Fletcher v. Experian Information Solutions, Inc., No. 25-20086 (5th Cir.), the tool identified every legal-authority error later cited by the court, the company said. It also found seven additional errors the court did not mention, including fabricated quotations, misstated holdings, and a citation that resolved to an entirely different case.

BriefCatch also pointed to a February 2026 opinion in the matter. Chief Judge Elrod found that counsel had used generative AI to draft the brief and "failed to check the brief for accuracy," BriefCatch said, and sanctioned counsel $2,500.

Guberman said the product would change quality control in litigation drafting.

"RealityCheck will help lawyers fix these errors before courts and opposing counsel find them," he said. "We want to improve the integrity of filings across the country."

Platform expansion

RealityCheck combines BriefCatch's editing product with citation-verification infrastructure powered by Counsel Stack. BriefCatch said the addition expands its platform from writing improvement into citation verification and AI risk management.

The launch follows recent corporate activity, including a Series A funding round and the acquisition of WordRake's patented editing technology. RealityCheck is available as part of the BriefCatch Next platform and as a standalone offering aimed at firms seeking citation verification.

BriefCatch is known for software that integrates with Microsoft Word and provides editing guidance for legal drafting. According to the company, its customers include law firms and courts. With RealityCheck, it is targeting what it describes as a growing compliance and reputational risk for litigators as courts increase scrutiny of AI-assisted work and expect lawyers to verify every cited authority and quotation before submission.