by Richard Finkelman
This article first appeared in Law.com on January 1, 2025.
2024 witnessed a seismic shift in the legal industry, driven by the rapid adoption of generative Artificial Intelligence (AI). Surprisingly, law firms, particularly those engaged in litigation, have spearheaded AI adoption, surpassing corporate legal departments, likely due to several factors, including the surge in post-pandemic data, the sector’s complex yet repeatable processes, and its substantial market size. Conservative law firms have traditionally demonstrated remarkable agility in embracing AI, often outpacing innovative corporations. While the US leads the global charge, the UK is just a little behind, with 96% of UK law firms adopting AI, resulting in a 43% increase in work quality and efficiency[i].
One of the primary concerns surrounding the use of AI in law is the phenomenon of “hallucination” – the generation of inaccurate or misleading information. While significant strides have been made to mitigate instances of hallucination, caution remains crucial. Studies indicate that even specialized legal AI models from leading providers like LexisNexis and Thomson Reuters can produce inaccurate or misleading information in 17% to 33% of queries[ii]. This, while an improvement over the higher hallucination rates observed in general-purpose Large Language Models (LLMs) on legal queries, remains a serious concern.
Law firms are increasingly adopting specific legal workflows that leverage Retrieval-Augmented Generation (RAG) and private models to mitigate this risk. RAG technology facilitates the creation of private models that minimize hallucinations and provide some assessment of answer accuracy. However, human oversight remains indispensable. Legal professionals must meticulously review and verify RAG model outputs to ensure accuracy and reliability. By maintaining robust quality control measures, legal professionals can effectively harness the power of AI while minimizing the potential for errors and misrepresentation.
Looking ahead to 2025, we expect several key trends to shape the legal technology landscape. Regulatory investigations involving Big Tech are likely to continue, and the Supreme Court’s 2024 overturning of the Chevron doctrine may lead to a surge in litigation as parties re-evaluate their chances of success in challenging agency interpretations. This new wave of litigation will test the limits of agency deference and establish new legal precedents.
We anticipate increased M&A activity as the new administration shifts towards a more business-friendly environment. This heightened M&A activity will inevitably lead to more investigations, disputes, and due diligence. Adopting AI technologies pre-deal will prove invaluable when navigating tight timelines.
While AI continues to advance, ethical concerns and the potential for misuse will remain paramount. Privacy will also be a key focus, with stricter regulations and heightened public awareness. A landmark legal case, potentially mirroring the significance of Judge Schira Scheindlin’s 2006 rulings on e-Discovery, could establish new legal standards for AI’s ethical and responsible use in legal practice. Such a case would have a profound impact beyond the legal profession, shaping the future of technology and its effects on society.
This underscores that 2025 will be a year of rapid innovation and complex challenges, demanding careful navigation and adaptation within the legal sector.
[i] https://www.lawsociety.org.uk/topics/business-management/partner-content/key-trends-shaping-the-legal-industry-in-2025
[ii] https://www.legaldive.com/news/legal-genai-tools-mislead-17-percent-of-time-stanford-HAI-hallucinations-incorrect-law-citations/717128/