Our Take
CiteSentinel solves a real problem (courts are sanctioning attorneys for AI hallucinations in citations) but is a detection tool, not prevention—the burden of review still falls on the lawyer who signs the filing.
Why it matters
Courts nationwide are imposing sanctions on attorneys who submit briefs containing invented case law, a documented failure mode of generative AI. Legal teams now face liability for citations they may not have written themselves, making verification infrastructure a professional necessity.
Do this week
Legal teams: audit your AI drafting workflow this week by running a CiteSentinel scan on one completed brief before final submission, so you baseline your current false-positive rate and liability exposure.
Courts sanction lawyers for AI-invented citations, CiteSentinel launches to catch them
BrentWorks, a legal technology startup, released CiteSentinel, a platform designed to detect fabricated or misattributed citations in legal documents. The tool scans case law, statutes, and legal authorities before filing and flags those that do not exist or are materially misstated.
Courts across the United States have begun sanctioning attorneys who submit briefs containing invented case citations, a known failure mode of large language models that generate authoritative-sounding but entirely fictional legal authority. CiteSentinel was built to close that verification gap. According to co-founder Brent Britton, an MIT-trained attorney and engineer, the tool "lets lawyers move fast with the efficiencies of generative AI while still filing documents reciting authorities they can stand behind."
The tool addresses a downstream liability problem: many attorneys who do not personally use AI to draft documents discover they have a citation problem anyway, because opposing counsel, co-counsel, contract attorneys, or paralegals may have used generative AI without disclosure. When a fabricated citation reaches the court, the name on the filing becomes the attorney's liability, regardless of who drafted it.
CiteSentinel allows users to scan three types of documents: their own AI-assisted drafts before filing, submissions from co-counsel and support staff, and opposing counsel filings for strategic advantage.
Biotech litigation faces outsized risk from AI hallucinations
The risk is especially acute in biotechnology and patent law, where cases depend on highly technical evidence including patent claims, prior art, clinical trial data, FDA regulatory history, and licensing agreements. An AI system could invent scientific references that do not exist, mischaracterize FDA guidance, fabricate patent precedents, or incorrectly summarize clinical trial results.
BrentWorks co-founder Brent Hunter, who applied neural networks to finance in 1993, notes that CiteSentinel is the first product in a series the company plans to release for legal practice in an AI-assisted environment. Britton frames the stakes directly: "In biotech litigation, where the technical record is everything, a ghost FDA guidance document or a fabricated prior art reference can unravel an entire legal strategy and years of work along with it."
How to use CiteSentinel in your workflow
The platform is built as a document scanner, not a drafting tool. It works after prose is generated, not during. This means adoption requires a new step in the document review process: running every brief containing AI-assisted text through verification before it goes to the printer or electronic filing system.
For teams already using generative AI in legal drafting, CiteSentinel functions as a gate before signature. For teams reviewing opposing counsel's filings, it provides a surface for identifying challengeable citations. The company does not publish false-positive or false-negative rates, so teams will need to baseline their own false-alarm burden during initial use.