My Experience Comparing Lexis and ChatGPT Deep Research
As a practicing attorney, I’ve been doing legal research for decades, starting with traditional platforms like Westlaw and Lexis, which I learned in law school and used throughout my early years in practice. But recently, I set out to compare those tools to something newer—ChatGPT’s Deep Research feature—in the context of an actual case I’m handling.
The Legal Question
The research question I explored was whether a utility company’s temporary, non-consensual entry onto private property to conduct a survey constitutes a taking under the Fifth Amendment and Maryland law. I used this same question across both platforms—Lexis and ChatGPT Deep Research—to evaluate how each handled it.
Using ChatGPT Deep Research
From the start, ChatGPT Deep Research was faster and more intuitive. I didn’t have to go through multiple login screens or navigate a cluttered interface. In fact, I could speak the question directly into the tool using voice-to-text—a feature Lexis doesn’t have—and it immediately began working.
ChatGPT didn’t just return a list of cases. It produced a comprehensive, reasoned memo that restated the question clearly and then walked through relevant federal and Maryland law. Most impressively, it identified the key U.S. Supreme Court case—Cedar Point Nursery v. Hassid—right at the start. That case is central to the issue of temporary physical takings, and Lexis completely missed it.
The memo didn’t stop there. It included an overview of other jurisdictions’ approaches, analyzed Maryland’s statute (Real Property §12-111), and even flagged a potential issue: that the utility company’s authority to enter the property might not fall under eminent domain. ChatGPT helped me think through legal nuances, even when I didn’t fully agree with its analysis—it gave me more to work with than any raw case dump.
Using Lexis
By comparison, Lexis returned 96 cases, along with hundreds of statutes and guidance documents—most of which were irrelevant. There was no memo, no summary, and certainly no effort to synthesize the law. Lexis didn’t ask me to clarify jurisdiction, nor did it attempt to understand the kind of takings claim I was exploring (physical vs. regulatory). I had to dig through pages of results and still didn’t get what I needed.
Cost Comparison
This project also made me reevaluate what I’m paying for. My Lexis subscription costs $270 a month, whether I use it or not. ChatGPT, on the other hand, costs me $200/month for enterprise access—and that covers all sorts of tasks beyond legal research. Even the $20/month plan would give most lawyers what they need for typical research. Plus, with ChatGPT, I can pause my subscription anytime. With Lexis, I’m locked in.
When Lexis Still Helps
To be fair, there are times when Lexis is still necessary—mainly when I can’t access a cited case or statute for free online. But now, instead of starting with Lexis, I’m using it as a backup to verify citations or grab a full-text decision I can’t get elsewhere.
A Shift in How I Practice
What surprised me the most is how much this experience has changed my approach to legal research. For years, I considered myself a legal research purist. I even just published a law review article that relied on traditional tools. But now, I find myself turning to generative AI tools like ChatGPT Deep Research first—not last. It’s not just about speed or ease of use. It’s about getting better, more usable results.
I genuinely hope law schools begin integrating tools like these into their curriculum. Not every lawyer will have access to expensive subscriptions, especially if they go solo or join a small firm. Knowing how to use tools like ChatGPT can make all the difference.
If you’re developing a legal tech product or want feedback on how to make your tools more lawyer-friendly, feel free to reach out to me through MyShingle.com or by email. I’d love to collaborate.