Analysis explores cost-effectiveness of AI and implications for future of legal outsourcing
A recent report has examined how the current capabilities of generative artificial intelligence (genAI) systems like retrieval and reasoning perform in comparison with senior and junior lawyers, legal process outsourcers, and specialised AI tools when finding and interpreting relevant information.
Litera, a legal technology software company, shared that its new report is tailored for legal professionals who want to stay ahead of the curve and to make informed decisions about the future of their practice.
Litera dubbed genAI as a potential game-changer in the rapidly evolving legal technological landscape. However, Litera noted that lawyers may end up overlooking the practicalities of integrating AI tools into their practice amid all the excitement that comes with this new technology.
The report provides a data-driven analysis and practical recommendations aiming to help lawyers evaluate and leverage genAI to improve the efficiency, accuracy, and profitability of their work. The report also gives insights on how to potentially enhance AI’s retrieval and reasoning capabilities, including by using hybrid search methods and advanced prompt engineering techniques.
Specific topics covered
Litera’s report titled “The State of Legal GenAI in 12 Charts” discusses how GPT-3.5 and GPT-4 compare with each other, as well as compare with traditional legal processes, in terms of performing key tasks relevant to legal practice.
The report also addresses the costs of using AI solutions vis à vis human labour and the consequent implications for the future of legal outsourcing. The report likewise tackles the strengths and weaknesses of genAI-powered legal tools, such as Lexis+ AI and Westlaw, in commercial applications.
Sam Fletcher, senior machine learning Engineer at Litera, noted that Australia does not have the strongest tech sector or the largest legal market in the world.
However, the invention of large language models (LLMs) has substantially reduced the barrier to entry for AI, Fletcher said. Often, jurisdictional differences can necessitate bespoke support in legal software, noted Fletcher, who obtained his PhD from Charles Sturt University in Australia.
Those interested in learning more about Litera’s report can access it here.