Skip to Main Content

Generative AI Tools for USD Law Faculty

Introduction

While people can interact with GPT-3.5 and GPT-4 directly, what will eventually be more common is interaction with GPT teamed up with other software. Lexis, Westlaw, and Bloomberg are all developing generative AI tools incorporating GPT while also adding layers both before and after GPT to adjust for accuracy and verification. These legal research generative AI products are designed for accuracy versus ChatGPT which is designed for language fluency. With these added layers, prompt engineering is less important, but knowing generally how AI works, as well as its strengths and limitations, will still benefit the user. 

Potential use cases (some of these perform better than others but since these products are still very young it's hard to say which):

  • Define legal concepts
  • Review documents/assist in eDiscovery
  • Generate summaries
  • Create deposition outlines 
  • Generate client letters, contracts, briefs, and patent applications
  • Identify relevant case law
  • Contract review and analysis 
  • Due diligence
  • Predict legal outcomes 
  • Basic legal advice chatbots
  • Compliance and regulatory monitoring 

More resources

  1. Andrew Perlman, The Implications of ChatGPT for Legal Services and Society, The Practice (Harvard Law School, March/April 2023) 
  2. LexisNexis, Generative AI for Lawyers: What It Is, How It Works, and Using It for Maximum Impact (May 2023)
  3. Cem Dilmegani, Generative AI Legal Use Cases & Examples in 2023, AI Multiple (May 2023)
  4. Jan Van Hoecke, What Are the Practical Use Cases for Generative AI in Legal?, Lawyer Monthly (May 2023)
  5. Mohar Chatterjee, How lawyers use AI, Politico (May 2023)

A note on hallucinations & quality of results

Legal research vendors have worked aggressively to build products that limit hallucinations and increase accuracy. First, most have developed specialized models trained on narrower, domain-specific datasets. The idea is that "good data," and only good data, is allowed in the system.  Secondly, most vendors are using retrieval-augmented generation (RAG) which takes the user’s question and passes it through a database that then adds to the user's question as “context” that is then sent through the model. Third, some products may also use vector embedding as a way to identify concepts, by way of assigning phrases or even entire documents, as numerical vectors.  Coupled with RAG, this increases precision and relevancy.  Lastly but certainly not least, almost all vendors incorporate human feedback on responses.  

It is true that these methods have reduced the risk of fake cases as seen in the Avianca case.  However these systems may still hallucinate with false sources or produce flawed outputs due to human user error, prompt misinterpretation, or missed issues. 

A 2024 benchmarking study indicated that legal generative AI models do reduce errors compared to general-purpose AI models like GPT-4, but that these models still hallucinate at an alarming rate. In fact, the Lexis+ AI system produced incorrect information more than 17% of the time, while Westlaw’s AI-Assisted Research and Ask Practical Law AI hallucinated more than 33% of the time.

Lexis+AI

Access: Lexis+AI is currently live and available to all U.S. firms and corporate partners.  In the academic space it is available at a small group of law schools with legal technology programs. USD faculty, administrators, 1L, 2L, and 3L students all currently have access to Lexis+AI.  Lexis consultants will be offer on campus training for faculty and students. Students should expect to see it in their summer associate positions though use terms and guidance will vary depending on the firm/institutions. 

Screenshot of Lexis+AI conversational search with link to Lexis+AI demo

Main features for the academic packages include:

  • Conversational Search
  • Document drafting* (e.g. draft me a cease and desist letter) 
  • Summarize a case

Document modifying (e.g. summarize this, make this more aggressive, make it shorter) and document upload are available in some law firm packages.  Pricing varies. 

Lexis+AI utilizes three LLMs (OpenAi, Anthropic (AWS), and Google Bert).  It is trained on both primary law (statutes, cases, constitutions, state and federal rules, and select administrative agency decisions and regulations) as well as a limited amount of secondary sources owned by Lexis (Matthew Bender, certain treatises, and Practical Guidance). 

Lexis has stated that Lexis+AI will not generate false sources.  Lexis+AI follows the RELX Responsible AI Principles to ensure accountability and prevent the creation or reinforcement of unfair bias. Both Lexis and Westlaw have stated that confidentiality is paramount and no user data will not be included in training data. 

Lexis' Generative Artificial Intelligence (AI) Resource Kit an overview of current practical guidance on generative artificial intelligence organized by practice area: IP, Data Security & Privacy, Civil Litigation, Commercial Transactions, Labor & Employment, Healthcare, Employee Benefits & Executive Compensation, Tax, and Federal Government. 

Become a Lexis+AI insider to stay abreast of current developments. 

Westlaw

Products and Access: As of January 1, 2025 all USD law faculty and students have access to Westlaw CoCounsel: Your Generative AI Legal Assistant.  Access Westlaw CoCounsel through the law school portal or Westlaw main landing page.Westlaw CoCounsel includes access to several different tools: 

  • CoCounsel Core
    • Review documents
    • Search a database of documents
    • Summarize law review articles and cases.
    • Contract policy compliance 
    • Compare documents 
    • Prepare for a deposition 
    • Timeline of events in a case
    • Draft (note that since CoCounsel is not fully integrated with Westlaw Precision's research capabilities, drafting from scratch is not recommended. Instead, one should draft using uploaded documents.)
  • Westlaw Precision with CoCounsel 
    • AI-Assisted research
    • Quick Check AI 
    • Jurisdictional survey AI - compare legal issues across multiple jurisdictions
    • Claims Explorer - identify potential claims based on fact patterns
  • Practical Law Dynamic Tool Set with CoCounsel 
    • Search and summarize within Practical Law
    • Additional Practical Law skills
  • CoCounsel Drafting 
    • Plugin for Microsoft Word to assist with finding, drafting, reviewing, and summarizing of existing documents. 

A fifth tool, HighQ Contract Express Legal Tracker is currently not available in our academic subscription. 

Westlaw account representatives will provide optional training in spring 2025.

Bloomberg

Bloomberg Law’s AI-powered tools launched Jan. 14, 2025.

  • AI Assistant, a chat-based tool that allows you to summarize and ask questions about documents you’re viewing, including Court Opinions, Court Rules, U.S. Code and CFR sections, State Statutes, Regulations, Session Laws, and Constitutions, as well as select secondary sources.
  • Bloomberg Law Answers, which provides a quick answer to your search question using relevant legal documents, including Federal and State Court Opinions, U.S. Code, CFR, State Statutes and Regulations, Practical Guidance, Bloomberg Law News, and secondary sources like books and manuals.

What's next?

Legal research behemoths Lexis, Westlaw, and Bloomberg are in an arms race to bring generative AI tools to their platforms.  For each of these companies the primary concern is "good data in, good data out."  

  • AI can only “learn” from the data it can access.  For Westlaw, Lexis, and Bloomberg this will mean they are constrained to things they own and things they have permission to load. It will take many months if not years to sort out permissions to law journal and news sources.  For obvious reasons sensitive company or customer data cannot be loaded. 
  • Bad training data can create bias or result in hallucinations.
  • After loading good data, who owns it? There is a risk that proprietary information could become part of open platform service if the information isn't walled off.. 
  • “Fine-tuning” the data going into the model, and its responses, is a challenge requiring many hours of human supervision. 
  • Regulators will need to emphasize the need for “explainable” AI. 

More resources on how vendors are fine-tuning their generative AI models: 

  1. Ken Crutchfield, How Generative AI Works (Part IV), Above the Law (August 2023).
  2. Sean Harrington, The Case for Large Language Model Optimism in Legal Research, SSRN (June 2023).
  3. How Do Domain-Specific Chatbots Work? An Overview of Retrieval Augmented Generation (RAG), Scriv.AI (August 2023).

The PDF below shows all the legal tech companies currently working on generative AI tools as of June 2024.