Jury Notes

Exploring Generative and Extractive AI in Legal Research: A Comparison of Leading Models

In preparing for my new class AI & Legal Research offered Spring 2025, I enrolled and completed the AI Prompting Certificate Course offered by UNH Professional Development & Training. The course helped me acquire the practical and strategic skills needed to excel in AI prompting. One of the "use cases" developed was using generative AI to create blog posts for business applications. With this in mind, I explored the use of ChatGPT produce posts for the Library blog. What are the implications of this practice? We decided to give it a try. I am examining and comparing the premium generative AI tools like Lexis AI and Westlaw CoCounsel. Lexis and Westlaw have been using extractive AI for years in their one search main search bar and other areas of the platforms. I used ChatGPT to produce the following blog post. Take a look and let me and Professor Zago know whether this is helpful and what you think of AI produced blog posts. I reviewed the post for inconsistencies with other sources I have consulted on the topic and I think ChatGPT did a good job. You may face questions like this in law practice, where firms large and small use blog posts to promote their practice areas.

Exploring Generative and Extractive AI in Legal Research: A Comparison of Leading Models

In the evolving landscape of legal research, artificial intelligence (AI) is playing an increasingly significant role. Models like Lexis AI, Westlaw CoCounsel, and others have revolutionized the way legal professionals access, analyze, and interpret vast quantities of legal data. This blog post will explore the distinctions between extractive and generative AI, compare some of the leading AI tools in legal research, and evaluate how these technologies are transforming the legal profession.

Understanding Extractive vs. Generative AI

Before diving into specific legal research tools, it is essential to define two foundational AI approaches: extractive AI and generative AI.

Extractive AI:

Extractive AI focuses on identifying and extracting specific pieces of information from a given dataset. For example, it can identify key phrases, sentences, or paragraphs from legal cases, statutes, or regulations. Tools leveraging extractive AI excel at pinpointing precise answers, such as locating a specific statutory provision or summarizing case holdings.

Key Features of Extractive AI:

• Pulls information directly from existing documents.
• Summarizes or highlights without creating new content.
• Excellent for citation-based research.

Generative AI:

Generative AI, on the other hand, creates new content based on its training data. It can draft documents, predict outcomes, or provide insights by synthesizing information. Tools built on generative AI models, such as OpenAI's GPT or Google's Bard, generate human-like text responses, offering interpretations, summaries, or even legal arguments.

Key Features of Generative AI:

• Creates new content rather than extracting pre-existing text.
• Produces nuanced responses by analyzing patterns in data.
• Useful for brainstorming legal arguments or drafting contracts.

Comparing Lexis AI, Westlaw CoCounsel, and Others

1. Lexis AI:

Lexis AI integrates extractive and generative capabilities to provide a comprehensive research experience. Leveraging machine learning, it assists users in:

• Conducting legal research with pinpoint accuracy.
• Summarizing case law and statutes.
• Drafting briefs and memos with its generative AI tools.

Strengths:

• User-friendly interface with robust search functionalities.
• Offers practical guidance and drafting assistance.
• Deep integration with the LexisNexis legal research ecosystem.

Potential Drawbacks:

• Heavily reliant on proprietary data, which may limit external source integration.
• Costs may be prohibitive for smaller firms.

2. Westlaw CoCounsel:

Westlaw CoCounsel, developed by Thomson Reuters, is another leader in legal research AI. It emphasizes advanced natural language processing (NLP) to:

• Answer complex legal questions.
• Perform due diligence and contract review.
• Enhance research accuracy using extractive AI techniques.

Strengths:

• Seamless integration with Westlaw's extensive legal database.
• Strong focus on practical application, such as contract analysis and litigation insights.
• Highly reliable citation tools.

Potential Drawbacks:

• Primarily extractive in focus, limiting its generative applications.
• Similar pricing challenges as Lexis AI.

3. Casetext (CoCounsel):

Casetext's CoCounsel is a notable generative AI-powered tool built on OpenAI's GPT models. It is designed to:

• Draft legal documents, such as discovery responses and demand letters.
• Conduct advanced legal research with conversational querying.
• Automate repetitive tasks like document review.

Strengths:

• Combines generative and extractive AI for diverse applications.
• Affordable for solo practitioners and smaller firms.
• Highly innovative in automating tedious tasks.

Potential Drawbacks:

• Limited access to proprietary databases compared to Lexis and Westlaw.
• Generative outputs may require additional validation for accuracy.

4. Harvey AI:

Harvey AI is another cutting-edge tool leveraging OpenAI's GPT-4. It has been adopted by several leading law firms to:

• Provide research insights through natural language queries.
• Draft legal documents.
• Generate summaries and arguments based on legal context.

Strengths:

• Designed for intuitive use by legal professionals.
• Focuses on streamlining workflows and reducing time spent on mundane tasks.

Potential Drawbacks:

• Still evolving, with limitations in specialized legal areas.

Key Takeaways: Extractive vs. Generative AI in Practice

Final Thoughts

AI models like Lexis AI, Westlaw CoCounsel, and Casetext CoCounsel represent the forefront of legal technology, blending extractive and generative capabilities to redefine what’s possible in legal research. As AI technology continues to advance, legal professionals must stay informed about its capabilities and limitations, ensuring they leverage the right tools for their specific needs.

By integrating AI effectively, legal professionals can enhance productivity, improve accuracy, and deliver better outcomes for their clients.