P2P

78 P E E R T O P E E R : I L T A ' S Q U A R T E R L Y M A G A Z I N E | W I N T E R 2 0 2 3 The Gen AI technologies receiving the greatest attention in the legal profession are LLMs (large language models). These systems are based on training deep neural networks on vast quantities of natural language text, enabling them to output remarkably fluent language in response to a prompt. For example, ChatGPT from Open AI combines an LLM with a chatbot interface trained to imitate the form of human conversations. This conversational interface enables almost anyone to experiment with the underlying LLM technology without the need to write software and has led to an explosion of interest in LLM applications. It is worth remembering, however, that LLM- based systems do not require conversational interfaces, and are likely to be increasingly embedded in software systems with more conventional software user interfaces. Applications of LLMs in the Law While attorneys are sometimes caricatured as slow to adopt technology, the law was a relatively early adopter of machine learning technology in the form of technology- assisted review (TAR) for electronic discovery (eDiscovery) in litigation, investigations, second requests, and similar applications. Unsurprisingly, a number of companies are investigating the use of generative AI to improve further search and classification capabilities in TAR systems, including reducing or eliminating the need to train TAR systems from examples. Entirely new capabilities are also being explored for TAR, including the use of Gen AI to answer questions, find supporting or contradictory statements within a set of documents, explore different uses of language within a document set, or generate prompt- directed summaries based on collections of documents. These new information access capabilities will have applications in the law far beyond TAR and discovery, in areas as diverse as researching case law, quickly understanding long and complex documents such as contracts, or coming up to speed on documents previously filed in ongoing litigation. The ability of LLMs to be prompted with background material, examples of desired outputs, and directions on how to modify previous outputs makes their capabilities for these tasks vastly more flexible than past technologies. Beyond their use in information access, these technologies also hold great promise for information management in contract management, legal hold, compliance and preservation, and information governance. However, the applications attracting the most excitement in the law are those that leverage the unique capabilities of LLMs to create coherent text. Much of law, and much of the time of attorneys, is concerned with the creation of documents: briefs, contracts, demand letters, advisory memoranda, and endless torrents of email communications with clients and others. Much of this content is highly stylized and repetitive, factors which encourage the possibility of automation. Further, law firms (particularly large ones) may be able to draw on masses of prior work product that could be used to tune general purpose LLMs to the particular needs of the firm. Challenges in Applying LLMs in Legal Work The combination of the enticing potential of Gen AI in document creation and the fact that conversational interfaces make experimentation with Gen AI available to a broad audience has led some commentators to predict very rapid changes in the practice of law. Despite the excitement; however, Gen AI is still a software technology and faces all the usual obstacles of integrating new technologies into legal work. A few moments remembering your firm's last integration of a significant new piece of software is a useful antidote to social media posts about how utopia or doom for attorneys is around the corner. F E A T U R E S

• Cover