Columbia Law School - July 28, 2024 - 8:12 am
Imagine a world where artificial intelligence (“AI”) doesn’t just assist lawyers, but actually speaks their language—a world where legal jargon and AI blend seamlessly, creating a symphony of digital jurisprudence. I believe this vision moved significantly closer to reality with the completion of a study I conducted with co-authors Aditya Sivakumar and Robert Simmons at UCLA School of Law.
Our study, titled “Standardized Nomenclature for Litigational Legal Prompting in Generative Language Models”, was published in Discover Artificial Intelligence on March 7, 2024. The study proposes legal prompting and communication techniques for lawyers. The study’s goal is to offer a framework that acts like a Rosetta Stone for lawyer-AI communication. The proposed system enhances the accuracy and reliability of AI-generated legal research, with the goal of changing the way legal professionals engage AI. You can access our study at: https://link.springer.com/article/10.1007/s44163-024-00108-5
A. The Framework
The new framework, essentially a syntax for legal AI interactions, includes:
1. Background variables/clauses: Essential facts and context of the legal matter.
2. Dependent clauses: Actions taken or desired outcomes based on the background information.
3. Generative clause: Instructions for the desired output from the AI.
4. Source clause: Requests for relevant legal sources and precedents.
5. Perspective clause: Assigns a specific role or viewpoint to the AI.
6. Tonality clause: Specifies the desired tone of the response (e.g., objective or persuasive).
“This standardized approach is crucial for maximizing accurate and consistent legal results from AI,” says author Aditya Sivakumar. “It’s about teaching lawyers how to speak ‘AI’ effectively.”
B. Putting AI to the Test
- Criminal Law: We tested the AI’s ability to analyze a case involving a fatal car accident during a police chase. The AI was prompted to consider charges ranging from first-degree murder to manslaughter, with and without requests for specific legal sources. When asked to cite sources, the AI provided more comprehensive definitions of crimes and cited relevant cases, demonstrating improved legal reasoning.
- Tort Law: The AI was tasked with building a case around a product liability scenarios. We found that the AI demonstrated an ability to identify relevant legal theories and cite applicable case law.
- Equal Pay: The AI was asked to analyze an equal pay scenario from both objective and subjective angles. We found that the AI showed an ability to shift its analysis based on a given perspective, demonstrating versatility in legal reasoning.
In each case, the AI demonstrated an impressive ability to cite relevant case law, outline potential legal strategies, and adapt its tone based on the given perspective. The results indicated that the AI’s responses were accurate, relevant, and aligned with established legal principles.
C. What This Means for the Legal World
- Efficiency Boost: Lawyers can now “speak AI” more effectively, potentially streamlining research and preliminary case assessments. The standardized prompts could save hours of initial research time and improve the efficiency of legal professionals.
- Enhanced Legal Outputs: The study demonstrated that using this standardized prompting framework significantly improved the quality of legal outputs from AI. The inclusion of source clauses ensured that the generated responses cited relevant legal precedents and statutes, enhancing the credibility and utility of the outputs.
- Versatility in Legal Reasoning: Perspective clauses allowed the AI to adopt specific viewpoints, such as that of a defense attorney or a prosecutor, providing more tailored and contextually appropriate advice.
- Ethical Considerations: The study emphasizes the need for responsible AI use, particularly regarding client confidentiality. Legal professionals must ensure that sensitive client information is handled appropriately and that AI-generated outputs do not compromise attorney-client privilege.
- Training Tool: Law firms and legal departments should train their teams on the use of standardized legal prompting. Understanding the components of the nomenclature and their application will enable legal professionals to maximize the benefits of AI.
D. Practical Steps for Implementation
- Adopt Standardized Prompts: Legal professionals should integrate the standardized nomenclature into their workflows, using the proposed framework to structure their interactions with AI. This will help ensure consistent and reliable outputs across different legal scenarios.
- Conduct Regular Reviews: Regularly review and update the prompting framework to incorporate new legal developments and technological advancements. This will help maintain the relevance and effectiveness of the standardized nomenclature.
E. The Road Ahead
While this standardization marks a significant step forward, I stress that this is just the beginning. Future studies could refine the nomenclature, explore its application in other legal domains, and assess its impact on the broader legal industry.
Some potential areas for future research include:
- Optimizing the framework for different areas of law (e.g., intellectual property, environmental law)
- Investigating the impact of AI-assisted legal research on case outcomes
- Exploring the ethical implications of AI use in law
As AI continues to make inroads into the legal profession, this new “legal prompting language” could be the key to unlocking its full potential while maintaining the accuracy and ethical standards the field demands. It’s a brave new world for law, where the next generation of legal minds might be as fluent in AI prompting as they are in legal statutes.
“We’re not replacing lawyers,” Sivakumar emphasizes. “We’re giving them a powerful new tool. But like any tool, it needs to be used correctly. That’s what this framework is all about.”
As this technology evolves, it raises intriguing questions about the future of legal education, practice, and even the nature of legal reasoning itself. The standardized nomenclature for legal prompting represents a significant advancement in using AI in law. Legal professionals can leverage this framework to improve their practice, ensure ethical standards, and navigate the complexities of legal scenarios with greater confidence.
One thing is clear, the gavel has dropped on a new era of AI-assisted law, and as AI technology continues to evolve, the legal industry must stay abreast of these developments, integrating innovative tools and practices to enhance the delivery of legal services. The future of law is here, and it speaks fluent AI.