vLex Research Assistant and Harvey AI
Lawyers are reminded almost daily that AI will transform the practice of law. Many have generally accepted this principle. One area where generative AI will significantly impact the daily practice of law is legal research and writing. Generative AI for legal research and writing is already making inroads in the legal field, assisting lawyers in various tasks like contract analysis, legal document generation, and more.
As demonstrated by two unfortunate New York lawyers and discussed in the MSBA blogs, "NY Lawyers and Law Firm Sanctioned for Citing Fake Cases" and "Using ChatGPT for Legal Research, Not so Fast!" the use of AI for legal research needs to be accurate (every single time!), helpful and dependable. The full acceptance and integration of AI generated legal research will depend on factors such as the technology’s accuracy, reliability, and the legal community’s willingness to adopt and adapt to these advancements. In practice, generative AI legal research tools must not hallucinate - ever! The tool needs to show real cases, statutes and regulations (authorities), it must be dependable and it must be up to date. In this article, the MSBA will review two separate AI legal tools and inform our members about their practical applications. Both of these tools are under development. One, vLex, uses its own legal library of authorities to produce relevant answers or legal documents in response to a lawyer’s query. The other, Harvey AI, uses its legal library and the internet to produce relevant answers or legal documents.
vLex® Research Assistant
vLex is a global legal intelligence platform that provides legal professionals with access to legal and regulatory information from its digital library consisting of a billion legal documents. It uses these billion legal documents to support generative AI legal research and writing.
vLex’s generative AI legal research tool is called Research Assistant. Currently, the jurisdictional legal content feeding Research Assistant consists of case law, statutes, and rules in the United States, the United Kingdom, Ireland and practice notes for Spain. They are working to expand the content sources to include more legal documents and jurisdictions. They are essentially taking the billion legal documents in their digital library and refining it to be helpful to lawyers. Their main priority in developing Research Assistant is accuracy and completeness over speed.
Research Assistant can conduct research on specific legal topics and draft a memorandum of law on the particular legal issue based upon the query asked by the lawyer. The memorandum of law will cite real legal authorities that are in the database such as statutes, rules, regulations, and cases (not hallucinated cases!) with hyperlinks to the citations. The results are based exclusively on authoritative materials in their legal database — not on third-party IP, the internet or hallucinations. Most importantly, if the assistant cannot provide an answer with confidence based upon the billion legal documents in the dataset, vLex will disclose and respond that they do not know the answer instead of creating an answer that is not in the dataset (knows what it doesn't know!). Thus, hallucinations are close to zero. According to vLex, it takes about two to three minutes per response to deliver a relevant memorandum of law, but they hope to reduce that. In comparison, two to three minutes is very small to the current way of conducting legal research on the popular legal research platforms and then manually drafting a memo.
In sum, what they have built is an intelligent assistant for question answering on legal topics. The lawyer can select or reject the research memorandum’s authorities which the assistance can use to produce more relevant answers. vLex’s Research Assistant prioritizes accuracy over speed. Legal issues a lawyer will be able to solve include, among others, pulling cases, summarizing documents, customizing summaries, writing memorandums of law with citations and hyperlinks to the case law and other authorities supporting the answer. Their vision is that the Research Assistant is AI that is “lawyer ready,” safe and reliable.
Harvey AI
Harvey AI is an AI platform specifically designed for the legal sector. It automates repetitive and mundane tasks, enabling legal professionals to focus on critical thinking, strategy and client communication.
Harvey AI is continuing to be developed to combine OpenAI’s GPT AI, which takes data off the internet, and combines it with legal case law (real cases) and reference materials. The vision is that by pulling information from authoritative legal sources and the internet, Harvey AI will provide a more focused and relevant result to legal professionals.
A handful of big law firms have endorsed Harvey AI as the next and best AI generative legal research tool for lawyers. Harvey AI is working with these law firms to develop custom tools, such as templates and legal drafting. When engaged by a firm, Harvey AI is trained using the firm’s own work products and templates to enable Harvey AI to provide tailored assistance to that specific firm.
Harvey AI’s capabilities include, among others, contract analysis, research, drafting, analysis and other routine legal tasks. Harvey AI leverages AI algorithms to analyze data, provide recommendations and make predictions to assist lawyers. Their tool works across multiple languages and practice areas. According to Harvey AI, a lawyer will be able to ask it to edit a legal document instead of manually editing the document. Most importantly, the makers of Harvey AI indicate that it can tell a lawyer what the differences are between an employee and independent contractor in a specific jurisdiction or ask it to rewrite a lease to comply with the jurisdictional requirements of a state.
Lawyers should stay in the driver’s seat
These products sound wonderful! They have the ability to conduct legal research finding relevant cases and other legal authorities. They can automate routine tasks and reduce the time it takes for a lawyer to conduct legal research or drafting, giving the lawyer more time to develop client relationships and concentrate on critical-thinking and strategy. Yet, these tools are not a substitution for a lawyer’s critical thinking and experience. A lawyer can take advantage of these AI tools, but stay in the driver’s seat! As AI continues to improve, it’s likely that generative legal research will become more widely accepted over the coming years.
Disclaimer: The MSBA does not endorse any of the above products or services identified in this article. This article is for informational purposes only.