Urgent.News

What's breaking now, across thousands of outlets.

AI

New online tool will check lawyers' sloppy AI-written briefs since they can't seem to be bothered

Bogus AI-generated citations keep turning up in court documents, but lawyers could at least run those docs through this legal citation checker

New online tool will check lawyers' sloppy AI-written briefs since they can't seem to be bothered

Lawyers who frequently employ AI to draft legal briefs face a significant obstacle: AI models often generate inaccurate case citations and misinterpret court opinions. A new tool called CiteProof aims to address this issue by meticulously examining briefs and filings for fabricated case citations and incorrectly quoted court opinions before they can lead to sanctions, fines, or embarrassing reprimands from judges.

This problem has persisted for years as lawyers have relied on AI for their work, with mixed results. While GPT-4 managed to pass a bar exam in 2023, it still suffered from hallucinations, as evidenced by Anthropic's lawyers mistakenly citing incorrect references during a copyright lawsuit defense.

CiteProof's spokesperson emphasized that the platform does not utilize AI models or probabilistic methods to detect AI hallucinations. Instead, it conducts direct and deterministic searches of public court records to verify the accuracy of the citations mentioned by AI-powered drafting tools. CiteProof asserts that their method does not transmit or store data on their servers, ensuring that all processing occurs within the United States.

The tool is free for individual checks and offers paid subscriptions for bulk scanning and more comprehensive reports. Although CiteProof cannot evaluate the quality of legal arguments, it serves as a useful safeguard to prevent the inclusion of fabricated citations in legal documents submitted to courts.

Written by urgent.news from The Register Science's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Also reported by 1 other outlet

Read the original at theregister.com →

More in AI

What it took to trust a package manager Claude wrote

I didn't write or even read a single line of code when making jpm, which is a package manager for JavaScript that works in npm, pnpm, yarn, and bun projects.

  • Claude created jpm package manager for npm, pnpm, yarn, and bun projects
  • Claude used AI to develop TLS, crypto, and http modules
  • Five days of testing ensured jpm's robustness across platforms

I Tried to Fix a Fuse on My Boat and Accidentally Invented a Marine AI Empire

I need to start with a confession: * this blog was never supposed to turn into whatever it is now. ** Truly. Honestly. I envisioned atmyboat.com as a fun, breezy chronicle of me learning to sail Our…

  • Accidental fuse issue on boat leads to AI experimentation
  • AAO-Methodology developed to control AI hallucinations
  • d3kOS marine operating system with AI features launched

What we learned giving AI agents their own mouse and keyboard

At Novastart ( https://novastart.com/ ) we built an infinite canvas OS: every desktop app lives on one endless canvas, and workspaces restore exactly as you left them.

  • Novastart developed an infinite canvas OS for desktop applications
  • AI agents assigned their own mouse and keyboard on the canvas
  • Lead agent organizes tasks, human pause feature ensures accuracy

More from Friday 16 October →