Can AI write a case law order? One judge decided to implement a unique idea to see how generative AI would handle three discovery disputes.

Last week for the Case Law Update session at Relativity Fest, moderator David Horrigan asked us all to identify our case of the year (so far). David identified the two rulings from 11th Circuit Judge Kevin Newsom where he used AI to inform his rulings here and here. I picked this case, which goes with (in my opinion) the issue of the year. Kelly Twigger picked this case, which was a great mobile spoliation case. And the Honorable Andrew J. Peck tapped into two of our favorite themes on the EDRM webinars we conduct together by selecting this case that dealt with 502(d) orders and this case that dealt with Judge Peck’s famous rule 1.1 (you’ll have to check out our webinars to find out what that is).

California Magistrate Judge Allison Goddard chose to do something a little bit different. With a list of cases to consider that we have covered so far this year, Judge Goddard had a “unique idea”. Instead of picking a case, she decided she would “take a page from my high school students and see if I could get some GenAI help with the assignment”. As she noted:

“I uploaded three cases into Claude.ai, gave it a basic set of background facts, and asked Claude to write an order using the principles set forth in the three cases.”

Here are Judge Goddard’s initial instructions to Claude (with the cases she uploaded linked for reference purposes):

I would like to combine the legal principles in each of these cases into one court order that I will use for demonstrative purposes only. The factual background for the order should be that there is a lawsuit between plaintiff and defendant, where plaintiff alleges that defendant sexually harassed her and she suffered emotional distress as a result. The first discovery dispute that the order needs to resolve is that Plaintiff served the following interrogatory on defendant: “State all times in the last 4 years that you have been accused of engaging in improper sexual conduct and identify the names and contact witnesses of all persons who have made such an accusation!’ Defendant objected that the request was overly broad, unduly burdensome, and invaded his privacy. He refused to answer the interrogatory. Plaintiff agreed during a meet and confer to limit the interrogatory to any other lawsuits or written pre-lawsuit demands accusing defendant of sexual harassment. Using the principles set forth in the Byte Fed v. Luxe Vending case, draft resolution to this dispute.

The second discovery dispute that the order needs to resolve is that Defendant accuses Plaintiff of selectively deleting her text messages. Defendant obtained text messages from one of plaintiff’s friends where plaintiff says that defendant is the perfect target because he has lots of money and her lawyer is going to make him pay for what he did. Plaintiff never produced any text messages, even though defendant served written discovery requests that would encompass text messages. Plaintiff stated that her phone is set to delete text messages automatically after 30 days and she did not change that setting after she filed this lawsuit. Defendant now seeks an adverse jury instruction and monetary sanctions against plaintiff for spoliation of ESI. Using the principles set forth in the Armstrong v. Holmes case, draft a resolution to this dispute.

The third discovery dispute that the order needs to resolve is that Defendant has refused to produce posts from his Facebook account, even though the court previously ordered him to do so. Defendant has not provided any valid excuse for this failure. Plaintiff now asks the court to compel Defendant to produce all of the posts from his facebook account for the past 4 years, without any redactions, directly to her attorney. Plaintiff also asks the court to compel defendant to produce all of the Facebook messenger messages for the same time period. Using the principles set forth in the Doe v. NCAA case, draft a resolution to this dispute.

After Claude generated an initial order, Judge Goddard responded, stating: “That was very good!” Can AI write a case law order? Hey, maybe it can!

She also asked it to update the plaintiff and defendant names – Sarah Connor and Inigo Montoya, respectively (which movie buffs will appreciate!) and asked it to “include references to federal rule of civil procedure 37 where appropriate. For the first issue, the analysis needs to address the proportionality required of federal rule of civil procedure 26.”

After the second iteration, Judge Goddard reviewed the result. While the proportionality reference to Rule 26(b)(1) was correct, the eight(!) references to various sections and subsections of Rule 37 were all over the board. So, Judge Goddard provided a third iteration of instructions, which said:

“I’m not sure you understand federal rule of civil procedure 37. Please go ahead and take those references out.”

To which Claude responded with: “I apologize for the confusion. You’re right, and I appreciate your feedback. I’ll remove the references to Federal Rule of Civil Procedure 37 as requested.”

Claude is certainly polite, at least! 😊

The resulting order is attached here. And it’s not bad! It’s short – just 4 pages to cover three disputes, so perhaps a bit too short. As Judge Goddard said when she emailed it to us: “It’s not the worst order I’ve ever read!”

High praise indeed! 😉

Nonetheless, this “unique idea” shows the possibilities of generative AI in helping judges write case law orders. Can AI write a case law order? Maybe not fully, but I could see it being a tool that could help judges develop ideas and a first draft. Idea generation is one of the best things generated by generative AI! 😀

So, what do you think? Do you think that judges should consider using generative AI to assist in writing case law orders? Please share any comments you might have or if you’d like to know more about a particular topic.

Disclaimer: The views represented herein are exclusively the views of the authors and speakers themselves, and do not necessarily represent the views held by my employer, my partners or my clients. eDiscovery Today is made available solely for educational purposes to provide general information about general eDiscovery principles and not to provide specific legal advice applicable to any particular circumstance. eDiscovery Today should not be used as a substitute for competent legal advice from a lawyer you have retained and who has agreed to represent you.