Generative AI is one of the most pertinent topics in today’s technological landscape, with game-changing implications for multiple sectors – particularly legal work. Firms and legal departments are increasingly showing interest in this technology, seeking to understand its potential benefits and drawbacks. Zach Warren from the Thomson Reuters Institute has provided some insights into the real-world applications of Generative AI, to help demystify the topic.
Legal firms have typically been cautious about adopting new technologies, but Generative AI is encountering a surprising level of enthusiasm. A survey conducted by Thomson Reuters revealed that while only 3% of firms had adopted Generative AI at the time of the survey, 82% of respondents believed it could be used for legal tasks, and 51% said it should be. This trend indicates that while many are still carefully observing the technology’s evolution, a significant chunk of legal professionals see potential in Generative AI.
As for overseeing its adoption, Warren suggests that responsibility primarily lies with CIOs, IT directors, and CTOs at present. However, he posits that partners and managing partners should also strive to understand this potentially transformative technology. In the long run, though, everyone has a stake in Generative AI, as it is designed to be implemented in every process, from onboarding new employees to day-to-day transactional work.
Warren comments on the potential for increased speed, efficiency, and consistency in research, document reviews, and repetitive tasks. However, this technology also raises questions about who will reap the benefits, notably the issue of what is billable in light of the new efficiency brought on by AI. Indeed, as 80% of corporate clients express their desire for their firms to use Generative AI, they also expect firms to prove their value, which brings an added emphasis on experienced lawyers.
What about the younger generation of legal professionals? Warren asserts Generative AI could revolutionize the work of first- and second-year associates, who traditionally handle a bulk of research and document drafting work. With a potential shift from legal writing to editing, skill development might soon become part of new legal recruits’ training.
However, it’s essential to discuss some concerns surrounding Generative AI, with ‘hallucinations’ as one significant issue. These are errors in AI output presented as facts, which the AI system can’t recognize as wrong as it’s not ‘intelligent’ as such but predicts the next word sequence based on previous results. Nonetheless, thanks to Retrieval Augmented Generation models, which learn from feedback about the accuracy of answers, this problem could be mitigated.
In the short term, the adoption of Generative AI should be approached with caution, especially when considering public, open-source tools. Warren suggests uses should, for now, be restricted to work where a margin of error is permissible, and verifying the output manually is essential.
To conclude, we are yet to witness the full impact of Generative AI in the legal sector, but it’s evident that it may play a crucial role in shaping future operations. However, it’s crucial that legal professionals remain mindful of both the potential benefits and the inherent pitfalls.