Shout Out: Carolyn Elefant's Review of Case Text v. ChatGPT!

/ The Tech-Savvy Lawyer.Page/

Here is a trimmed down version of Carolyn’s presentation. The original can be found on loom.

Previous podcast guest Carolyn Elefant of MyShingle.com reviews the AI legal research capabilities of Casetext and ChatGPT. She highlights Casetext's robustness in providing specific cases, while recognizing ChatGPT's speed and ability to give a broader overview. Despite favoring Casetext, she notes both tools as valuable for different aspects of legal research.

In my opinion, I would start with ChatGTP to give you a broad understanding of the issue.  Then, I would start using Casetext to help narrow down the issue and perhaps get some good research citations.  Regardless of how you use AI in your research, check your work, case citations, etc.  Don't be like the New York attorney sanctioned by the court for blindly using AI.

CHECK YOUR WORK!!!

Transcript:

Hi there, it's Carolyn Elephant of my shingle. Welcome to this SmackDown Comparison of the artificial intelligence legal research capabilities of Casetext and ChatGPT. Probably know these companies have been in the news recently, Casetext because it was acquired for a whopping $650 million by Thompson Reuters p t because it produced.

Hallucinations for an attorney who used it to research a case resulting in his being sanctioned for relying on fake cases. Sure. Before we start, I am a huge case text fan. I've been using the product for seven or eight years, and I also have an interest in the company for this research. SmackDown, I chose a question that has also been something that's been in the news.

A case about a guy in New Jersey who sent contractors to his neighbor's property to remove 32 large trees so that he could have a better view of Manhattan at sunrise or sunset. That, of course, triggered a lot of penalties or liabilities, so I thought we would ask Chachi PT about some of the consequences of unauthorized tree removal in New Jersey.

Going to start with case tech's, flagship product co-counsel.

We're going to start with case texts, flagship product co-counsel, and I'm going to say that I want to get some legal research on

tree removal, and I'm gonna select the legal research memo option. And then I'm going to put in my query. I represent a client who chopped down 32 trees on his neighbor's property. Without permission, what kind of damages or penalties will my client be responsible for? I'm gonna submit my query,

just give co-counsel a second. So co-counsel has interpreted my query. It also gives me an option to suggest what jurisdiction I want to apply. So I'm gonna go. Down and ask it to apply New Jersey case law and I'll click applied jurisdiction. I'll include on published decisions, no date filter, and I'm gonna submit my query and we'll just let it have a work at that.

So here's the result. This took, I would say about three minutes. And the answer is, um, my client may be liable for damages. Chopping down the tree and measure will depend on the facts of the case. It could either be diminution of value of the property or cost of restoring the property to its original condition.

And then p punitive damages can be awarded also. So then it lists 13 cases. And then case text gives me a little blurb about what the case is about. So here it says, This case discussed different methods for assessing damage for injury to land, including diminution of value or restoration costs. So this sounds like a case that I'm really going to want to focus on.

And of course, I do want to. Read the case. I don't know. This came up with an error. Now I'm not sure why that is, but I am going to want to read through the cases that I actually use. I would not just rely on the cases or even these blurbs here. Um, even though the summaries are pretty good, I'd still go in and take a look at the cases.

And one of the things the case text cases did not appear to cover is something that was mentioned in the article, which is the possibility of municipal citations or even trespassing. So I'm going to go back in and now follow up with case text and ask it will my client face liability under municipal ordinances for unauthorized tree removal?

So let's add that query here. So, And I wanted to do a legal research memo again.

Let me submit that, let it get to work.

Once again, I'm gonna plug in the filter that I want for New Jersey.

And submit the request and let it get to work.

And here's the response That case text came up. It says that municipal ordinances can make a property owner liable. There's a case here that addresses whether Muni can help somebody libel, um, for unauthorized tree removal, discusses a statute that prohibits local governments from requiring a permit for tree removal.

So there do seem to be some discussions of cases, but I think the important point here is that. Any research you do, no matter how robust the tool is, not necessarily always going to give you the responses you want. That's not a deficiency in the legal research tool. It's just that legal research is an iterative process.

And so in this particular situation, my first inquiry didn't answer a question or didn't respond to an issue that I saw in the article about municipal LI liability under municipal ordinances. So now we're gonna go ahead and take, I look at the competitor ChatGPT. I use ChatGPT plus, so that just makes sure, ensures that I have access all the time and I'm putting in the identical inquiry.

I represent a client who chopped down 32 trees. What kind of damages are penalties? Let's see what G P T four says. Now, this is a lot faster, as you can see. But as you can also see, it's not as robust, but Chachi PT does pick up on the main issues. It says that there could be civil liability, the value of the trees, economic loss for removal, additional compensation for environmental harm, restitution to the neighbor, cost of replacing the trees, maybe statutory damages, fines, penalties, and injunctions.

So it actually chachi PT in response to that question, even though its response is much less. In depth on these topics. It gives a broader overview and captured all of those issues, including statutory penalties and municipal liability on the first run, which at G P T I'm gonna follow up with a question about, can you cite me any cases under New Jersey law?

And of course, we all know that we operate at our peril when we ask it to psych cases. Let's see what it says.

Oops.

And so here, ChatGPT tells me that it can't cite specific, uh, case citations and it refers me to Lexi Nexus or Westlaw. It doesn't refer me to, uh, to case texts, but it does refer me to other resources. So from my perspective, case text is hands down, the better tool. It's more. Robust. It gives me cases that I can research and follow up with.

ChatGPT though is not worthless. It certainly gives a lay of the land and is at least starting point or a springboard for future research, and it is a lot faster. If there are research tools that you want me to try another SmackDown with, let me know. You can reach elefant@myshingle.com. Hope you found this useful.