Dickerson: A review of leading AI programs in the legal industry

The jury is back and has rendered its verdict — generative artificial intelligence in the legal industry is here to stay. Despite a litany of concerns with integrating AI into the legal industry, corporate legal departments, government attorneys, law firms and even the judiciary are embracing the use of AI at exceeding rates. For better or worse, every actor in the legal system has unique incentives for adopting AI. Government agencies at all levels see potential for AI to address limited resources and increasing workloads. Law firms and corporate legal departments are already employing AI programs for repetitive tasks such as contract and document review. Courts are even exploring AI for workforce management, processing information and analytics to predict the outcome of cases. While everyone is seemingly exploring the use of AI, not all programs are made alike. This article will review legal AI programs and how legal actors are diving headfirst into the brave new world of AI and the law.

Law firms and corporate legal departments have predictably been the first legal actors to capitalize on the use of this magical technology. While there are clear efficiency and competitive gains to using AI, it seems that partners and chief legal officers recognize that if they don’t lean into AI, their clients will bypass their advice and create even bigger risks. Where they deviate, however, is their choice to use an outside vendor or develop an internal AI program. Well-resourced firms such as Dentons (fleetAI) and Addleshaw Goddard (AGPT) are launching proprietary versions of ChatGPT for legal research, content generation and document analysis. Firms and legal departments are balancing customization and control against cost, privacy and reliance on vendors. These proprietary legal AI programs are in their infancy, and only time will tell how they match up to their third-party competitors.

Firms and consultancies such as Allen & Overy and PwC have adopted the ChatGPT-style platform Harvey. Harvey is an AI-driven legal research and analytics tool that aims to revolutionize the way legal professionals conduct legal research and gain insights into legal matters. It uses natural language processing to understand and interpret legal documents. It also provides robust search capabilities, allowing users to efficiently sift through vast legal databases to identify pertinent information and potential issues. Unlike the free version of ChatGPT, Harvey can evaluate the authority and relevance of cited cases. Similar to Westlaw and LexisNexis, it can also offer insights and data visualization tools that enable users to track legal trends, monitor changes in caselaw and gain a deeper understanding of legal issues. Harvey isn’t entering the market as an ancillary tool — it is designed to be the foundation of a firm’s workflow to make informed decisions and shape legal strategies.

Attorneys are likely more familiar with Westlaw and LexisNexis. Both have entered the marketplace with their own versions of chatbot-style AI programs. In August, Thomson Reuters Corporation announced that it had acquired Casetext Inc. for $650 million in cash. As a trusted leader in legal research, Thomson Reuters decided to build on existing technology rather than develop its own. Founded in 2013, Casetext was among the first to use advanced AI and machine learning to assist attorneys with efficiency. Its key product, CoCounsel, is powered by GPT-4 and can perform document review and contract analysis, draft legal research memos, and prepare an attorney for a deposition. Thomson Reuters also plans to leverage Microsoft’s CoPilot AI to assist users with drafting legal documents. This is a more comprehensive approach to adopting AI programs and signals that companies will continue to expand the scope of these programs to keep up with demand.

LexisNexis has entered the AI space, albeit with less noise than some of its competitors. With its AI program, Lexis leaned into analytics and marketing as its chief advantage. The program boasts an impressive analytics platform built to find and target new clients; identify insights into competitors’ strategies; distill copious data into a clear visualization tool; and predict case outcomes. Lexis has pitched its AI to cull all the tools together to develop a comprehensive case strategy, from finding clients, proposing venues and judges, and shaping strategy based on expected outcomes. Unlike other programs, Lex Machina offers a degree of transparency by allowing users to generate customized reports that show underlying dockets, documents and trends to understand where the information comes from.

The list of legal AI programs goes on, and perhaps users will eventually need a chatbot to weigh the pros and cons of each one. Each program has unique features, and the enterprise has the potential to revolutionize the legal profession. This may explain why so many legal actors have jumped at the opportunities that AI has to offer faster than any other technological development to hit the industry. Many commentators have astutely identified concerns with the ethical considerations, security and privacy, and bias. But arguably the biggest pitfall in any of these programs is their questionable accuracy.

Simply put, legal AI models are trained on data, and the quality of the data can affect the accuracy of the model. If the data is biased or incomplete, the model will be biased or incomplete, as well. We’ve already seen how embarrassing it is to rely on the inaccurate information generated by ChatGPT and how devastating it can be for a client to lose their case based on those mistakes. Above all else, clients and citizens want legal deliverables to be accurate, because whether through invoices or taxes, we’re all paying for this technology in some fashion. These programs are still very costly and not yet scalable, making the potential mistakes all the more expensive. Some practitioners will use this point to champion the old school ways of practicing law. Others will see a competitive advantage and invest in platforms that align with their specific needs to deliver better services. Consumers of legal services — whether they are hiring elite global law firms or the veteran attorney on the town square — will soon reflect their preferences. But as of now, there seems to be an insatiable desire to see the full potential of this technology. •

__________

Conner R. Dickerson is a member of the Business Services, Real Estate, and Business Litigation practice group at Cohen & Malad LLP. Opinions expressed are those of the author.