Panoram recently hosted a roundtable event discussing the impact of ChatGPT on the legal industry. This exclusive gathering brought together senior legal and governance executives to explore the implications of this disruptive technology. Notably, ChatGPT was present as a VIP guest, engaging through voice commands and showcasing its capabilities. Here’s a summary of the key takeaways from the discussion:
ChatGPT where are we now?
It was clear to see, in the ever-evolving landscape of artificial intelligence, ChatGPT has emerged as a topic of interest for the legal industry and our delegates were keen to learn more about this technology. With its training on a vast array of legal documents and knowledge, our delegates agreed that ChatGPT has developed a good understanding of legal rules and procedures. Some delegates suggested it excels in some tasks more than others such as interpreting and applying specific arbitration rules, including determining the type of arbitration required and the number of arbitrators involved. Others compared ChatGPT’s capabilities to ‘AutoFill’ making it a valuable but limited asset for generating generic legal content, saving time and effort for legal professionals.
Risks and limitations
Whilst recognising the benefits of ChatGPT, our delegates raised valid concerns during the roundtable. Notably, an insurance representative highlighted the challenge surrounding professional indemnity insurance, emphasising the potential liabilities and claims against legal professionals and firms arising from inaccurate or flawed legal advice provided by ChatGPT.
Training emerged as another crucial concern. As the legal industry incorporates ChatGPT into daily working life, comprehensive training becomes imperative to ensure employees understand the limitations and potential pitfalls of relying on AI-generated content. Without sufficient training and fact-checking mechanisms, legal professionals may overestimate ChatGPT’s capabilities, leading to undue reliance on its outputs without critical analysis, potentially resulting in errors, misinterpretations, and inadequate legal advice. Delegates also discussed the need for evolving skill sets, emphasising the demand for professionals with a deep understanding of the technology or “Prompt Engineers” who can effectively craft prompts for optimal outputs. There was also some apprehension around the table over junior lawyers having less exposure to learning through manual document review that could now be carried out by technology.
Compliance issues were also highlighted, as ChatGPT processes and stores user interactions, posing risks of unauthorised access, data breaches, and inadvertent disclosure of sensitive information.
Lastly, the risk of flawed or biased data being fed into ChatGPT was emphasised, which could undermine the reliability and integrity of the model’s outputs, jeopardising the accuracy and fairness of legal advice. Several of our delegates also questioned what would happen when AI creates more content than humans. Will it reach a limit of learning?
What does the future look like?
Delegates acknowledged the intriguing future prospects of generative AI in the legal industry and beyond. Predicting case outcomes by analysing legal precedents and relevant data was one prominent possibility discussed, potentially reshaping the roles of judges and traditional litigation processes.
Furthermore, in-house legal teams are expected to increasingly rely on ChatGPT to enhance efficiency and self-sufficiency. By leveraging the model’s capabilities, they can draft contracts, conduct legal research, and generate basic legal documents autonomously, reducing reliance on external counsel, cutting costs, and improving turnaround times.
Larger law firms may explore building proprietary AI models using their extensive internal data, leading to a potential polarisation within the industry. Some firms could gain a competitive advantage through specialised and customised AI tools exclusive to their organisation.
Additionally, delegates agreed that as AI technology progresses, new competitors similar to ChatGPT will likely emerge, fostering innovation, competition, and the development of specialised models tailored to specific legal domains.
During the event, ChatGPT was tested alongside one of our expert delegates, who provided his own answers for comparison with ChatGPT’s output. The comparison demonstrated that while ChatGPT is a useful productivity tool, it currently does not replace the expertise, experience, and trust that professionals offer. It should be viewed as an aid to enhance productivity, with professionals possessing the knowledge and ability to question and fact-check the generated output. That’s where it is today, query where it will be tomorrow given its recent exponential improvement.
To further explore ChatGPT’s impact on the legal industry, you can refer to our previous blog, where we asked ChatGPT to summarise its transformative role, and compare it to the insights shared by our delegates.
Panoram will be running a series of roundtables on this topic. If you would like to join us, please email our Head of Marketing, Rebecca Tear – [email protected]