David Wakeling, who heads the markets innovation group at Allen & Overy, a London-based law firm, was introduced to Harvey, a law-focused generative AI tool developed by OpenAI, in September 2022. Initially, he initiated a small experiment wherein a few lawyers from his firm used Harvey to handle basic legal inquiries, draft documents, and create initial client communications.
The trial started modestly but quickly expanded in scale. Approximately 3,500 employees spread across the firm’s 43 offices began utilizing the tool, generating approximately 40,000 queries in total. As a result, the law firm has forged a partnership to extend the use of this AI tool throughout the organization, although the specific value of the agreement remains undisclosed. According to Harvey, one out of every four lawyers at Allen & Overy now incorporates the AI platform into their daily workflow, with 80 percent using it at least once a month. This trend is also catching on in other prominent law firms.
The legal industry has witnessed predictions of AI disrupting its operations in the past, but the emergence of the latest wave of generative AI tools, with ChatGPT leading the way, has instilled greater confidence within the industry.
David Wakeling views this as the start of a significant paradigm shift, particularly emphasizing the compatibility of this technology with the legal field. Generative AI, which leverages extensive datasets to produce natural-looking text or images, aligns well with the legal sector’s reliance on standardized documents and precedents.
Lilian Edwards, a professor of law, innovation, and society at Newcastle University, believes that legal applications like contract drafting, conveyancing, and license generation are well-suited for ChatGPT and similar technologies. Legal document generation has been a growth area for decades, and even in the era of rule-based tech, law firms heavily relied on standardized templates and precedent banks, resulting in more predictable outcomes compared to free-text outputs.
However, current generations of generative AI have demonstrated some issues, most notably their tendency to generate inaccurate information, also known as “hallucinations.” While Harvey’s founder and CEO, Gabriel Pereyra, assures that they have implemented measures to minimize hallucinations in legal use cases, mistakes have still occurred, necessitating careful risk management.
Users accessing Harvey at Allen & Overy are presented with a set of rules, with the most crucial one being the validation of all output from the system to ensure accuracy.
Although some lawyers are cautiously optimistic about the use of AI in their practice, there is recognition that AI may excel in handling entry-level tasks, such as drafting legal documents. However, the need for thorough review remains paramount, as the nuances of each client’s circumstances often require a human touch.
Additionally, the management of data inputs into AI systems poses its own challenges, particularly regarding data privacy and confidentiality. In Europe, the use of AI that involves personal data may raise concerns about compliance with the General Data Protection Regulation (GDPR).
Allen & Overy emphasizes its commitment to safeguarding client data and ensures that it is not used for its deployment of Harvey unless stringent protections are in place. As international regulations tighten regarding the use of personal data with generative AI tools, law firms like Allen & Overy are working to strike a balance between utilizing AI to enhance productivity while upholding data privacy and security.
Overall, David Wakeling believes that AI will bring tangible improvements to productivity and efficiency within the legal industry, even if it doesn’t completely disrupt traditional practices. Small tasks that were once time-consuming can now be outsourced to AI, making a significant impact when scaled across thousands of lawyers within the organization.
Leave a comment