Many probably know of NotebookLM as the AI collaborator where you can upload documents (“sources”) to a digital “notebook” and then ask questions about the information in your sources. NotebookLM responds with an answer and citations to the sources you’ve supplied. Since you control what NotebookLM focuses on, there’s less worry about made-up hallucinations. NotebookLM…
Tag: AI
Four different prompts. Asking Claude.ai to help understand a divided appellate decision
Three justices wrote when the Michigan Supreme Court released Thursday’s 62-page decision in Shareef El-Jamaly v Kirco Manix Const (164902-4). Justice Welch penned the 32-page majority (joined by Justices Bernstein, Cavanagh, and Bolden). Chief Justice Clement filed a 3-page opinion that concurred and dissented in part. Justice Zahra authored a 27-page dissent (joined by Justice…
When 4 jurists write: Prompting Claude.ai to create tables of disagreement and agreement and asking follow-ups in the ex parte decision People v Loew (164133)
Four justices wrote when the Michigan Supreme Court released today’s 93-page decision in People v Loew (164133). Chief Justice Clement penned the 29-page lead majority (joined by Justices Zahra and Viviano). Justice Welch filed a 31-page opinion that concurred and dissented in part (joined by Justice Cavanagh). Justice Bolden wrote for herself in a 15-page…
Prompting Claude.ai to create tables of disagreement and agreement in the indigent/expert-witness decision People v Warner (163805)
A divided Michigan Supreme Court released its 39-page decision in People v Warner (163805) last Thursday. I deleted the Syllabus pages and put Claude.ai (a next-generation AI assistant) to work on the majority and dissenting opinions by creating tables showing the areas of disagreement and agreement. Here is the prompt: Create a three-column table that outlines the key…
Using Claude.ai to understand the package-deal plea withdrawal decision People v Samuels (164050)
A divided Michigan Supreme Court released its 32-page decision in People v Samuels (164050) on Friday. I deleted the Syllabus pages and put Claude.ai (a next-generation AI assistant) to work on the majority and dissenting opinions. First, I asked Claudi.ai to outline the decision using this complex prompt: Draft an opinion outline of the attached…
Prompting Claude.ai and ChatGPT to outline a new Michigan Supreme Court decision
Today, the Michigan Supreme Court issued a split opinion in People v Prude (Docket 165664)—a case that was decided without oral argument. Because users can now attach files with their “prompts” when using the free versions of the generative AI tools Claude.ai and ChatGPT, I decided to take each for a spin to outline the…
Protective orders that prohibit submitting confidential discovery material to open generative AI tools
In a sign of the AI times, the federal court in the Southern District of New York recently entered a protective order that includes an AI-restriction paragraph:
Asking Claude.ai to outline a court decision
Last August, I shared my experience using Claude.ai to outline case briefs (pretty good!). Today, I took the current (free) version of Claude.ai for a spin to outline a recent unanimous Michigan Supreme Court opinion. tl;dr: Claude.ai did a really good job. My only quibble is that it left out the role of one of…
How judges can draft and add easy-to-understand summaries in their written decisions
Summary for Pro Se Plaintiff The magistrate judge is recommending that the motion for summary judgment be granted. Your evidence does not show that Nurse Yule, Nurse Practitioner Housley, or Dr. Wong acted in ways that they knew were likely to cause you additional harm or suffering. Such evidence is required to show a violation…
Procedural takeaways when there are AI/fake-case concerns
In less than a year after the ChatGPT usage in Mata v Avianca, Inc., SD NY (Docket No. 1:22-cv-01461) captured headlines, we’re starting to see a pattern and lessons for how adverse parties and courts are handling suspected fake-case court filings. The pattern often follows the routine of other claims when reliability may be questioned…