By Brendan Ford, Founding Partner
Artificial intelligence has been swiftly reshaping the legal industry, and I’ve witnessed its impact on legal research. Integrating AI-powered tools has revolutionized document review, legal analysis, and drafting tasks. It’s important, though, to acknowledge the existing constraints of AI in legal research, which both lawyers and clients need to consider. In the following piece, I’ll share my journey into AI legal research, shedding light on its transformative potential.
In a recent case, a New York lawyer was questioned for two hours by a United States District Court Judge for using an AI-powered tool to create a legal work product without double-checking its accuracy. According to the New York Times article, “the lawyer had created a legal brief for a federal-district court case filled with fake judicial opinions and legal citations, all generated by ChatGPT.”
After reading this article, I had trouble believing that AI would generate fake judicial opinions and legal citations. Instead, I assumed that the citations would be near the mark but not spot-on due to AI’s limitations in understanding the nuances of the law. The law is complex and often requires the ability to interpret and apply precedent, which AI systems are still struggling to do. For example, an AI system may be able to identify relevant legal authorities. Still, it may not be able to understand the nuances of those authorities or how they apply to the specific facts of a case. I did some AI-generated “legal research” to test these theories. The results were surprising.
I had a recent case involving a business dispute. One of the parties entered into contracts before formally becoming a corporation, yet sued other parties for allegedly breaching those contracts. The legal issue was straightforward: can a nonexistent corporation enter into binding contracts, and under what circumstances?
I asked Google Bard for help. I picked it because Google Scholar has countless cases in its database, and I assumed that it would have the best shot of giving me accurate responses.
My Prompt:
“Send me 3-5 California case state citations for the proposition that a nonexistent corporation can enter into binding contracts.”
Bard’s Answer: (verbatim):
Sure, here are 3-5 California state case citations for the proposition that a nonexistent corporation can enter into binding contracts:
- Carrott v. C.C.C. Enterprises, Inc., 24 Cal. App. 4th 1073 (1994)
- John Wiley & Sons, Inc. v. Livingston Enterprises, Inc., 216 Cal. App. 3d 1040 (1990)
- California Pacific Bank v. Universal Financial Services, Inc., 214 Cal. App. 3d 83 (1989)
- In re American Construction Co., 152 Cal. App. 2d 181 (1957)
In Carrott v. C.C.C. Enterprises, Inc., the court held that a nonexistent corporation could enter into a binding contract if the parties to the agreement intended to be bound by the contract and the corporation subsequently came into existence.
In John Wiley & Sons, Inc. v. Livingston Enterprises, Inc., the court held that a nonexistent corporation can enter into a binding contract if the party with whom it contracts knows that the corporation is not yet in existence but intends to form it.
In California Pacific Bank v. Universal Financial Services, Inc., the court held that a nonexistent corporation can enter into a binding contract if the party with whom it contracts is aware of its existence and has reason to believe that the corporation will come into existence.
In In re American Construction Co., the court held that a nonexistent corporation can enter into a binding contract if the party it contracts is unaware of the corporation’s nonexistence.
It is important to note that these cases are all based on California law. The laws of other states may vary.
Eureka! Bard found four California state cases that told me if and when a nonexistent corporation can enter into binding contracts — precisely what I wanted. Now I need to check to see if these are still good cases, and I’ll be good to go, right?
Well … not exactly.
I pulled up the first citation: 24 Cal. App. 4th 1073 (1994).
Hmm … it’s not Carrott v. C.C.C. Enterprises, Inc., nor is it a citation to the beginning of a case. Nor does the case have anything to do with nonexistent corporations and contracts. The citation must be wrong; I’ll look under the case name. After spending a bit of time on this, I came to a harrowing realization: there is NO REPORTED CASE ANYWHERE IN California with that name. None. It’s a totally fabricated case.
This had to be a one-off mistake, right? Wrong.
John Wiley & Sons, Inc. v. Livingston Enterprises, Inc., 216 Cal. App. 3d 1040 (1990) doesn’t exist either. The case citation is inaccurate. And the only case name that matches is a United States Supreme Court case from 1964 involving collective bargaining — nothing to do with nonexistent corporations and contracts.
The other two cited cases were just as bad. The citations didn’t match a case. The case name was wrong. And the cases had nothing to do with my research. Bard, with full access to Google Scholar, not only failed to provide cases to help my assignment but provided me with fake case law. I would have been in profound trouble if I had mindlessly cut-and-paste the answer into a brief.
AI is a powerful tool that has the potential to revolutionize legal research. However, lawyers and their clients must be aware of AI’s limitations and use these tools responsibly.
Here are three tips for using AI for legal research responsibly:
- Be aware of the limitations of AI. AI systems are only as good as the data they are trained on and the algorithms used to train them.
- Use AI tools in conjunction with human judgment. AI tools should not be used as a substitute for human judgment and expertise.
- DO NOT CUT-AND-PASTE AI answers into documents. Check, double-check, and triple-check to ensure that references, facts, statistics, and other factual representations are 100% accurate.
The Takeaway:
Every company and business leader must recognize the significance of the AI revolution. Implementing AI policies for employees is imperative to ensure legal compliance, address ethical concerns, and effectively integrate AI technologies. Contact us to take action now to safeguard your business and promote responsible AI usage.