From the DRI Center for Law and Public Policy AI Working Group
Artificial Intelligence (AI) has become an increasingly important topic in the legal industry in the last few years following the introduction of ChatGPT and other generative systems. But AI is not new to the legal industry. In fact, there are many legal support programs that use AI to improve functionality and efficiency. Any system that makes decisions autonomously—as opposed to being human operated—is AI. It has been incorporated into legal research platforms, document review programs, e-discovery systems, Microsoft Word and Outlook, and yes, chatbots like ChatGPT. While AI has come a long way, it is continuing to develop, and it is by no means a replacement for legal analysis and reasoning employed by experienced lawyers. Rather, it should be seen as a tool to be used by lawyers to perform tasks with a higher level of efficiency and quality.
Because AI is not a replacement for legal training and experience, lawyers must remember that our ethical obligations to our clients and the courts require that we review any work product from AI for completeness and accuracy. Additionally, lawyers must carefully consider the confidentiality and privacy concerns that come with the use of AI. Text generators and chatbots like ChatGPT initially did not have any privacy restrictions for information entered into the system. Developers have started to institute confidentiality restrictions for these systems, including building generative systems that are exclusive to the law firm for which the system is developed. However, even when exclusive to a law firm, these systems are designed to use the information entered by users to learn and respond to future questions from users. Similarly, there are significant questions about who owns the work product from AI, further complicating the confidentiality and privacy issues. The ABA has started to analyze the confidentiality and privacy questions posed by AI and has provided some initial guidance. Lawyers using AI must carefully consider what information can appropriately be entered into these systems without deviating from our obligation to keep our clients’ information confidential.
AI is also not immune from bias. AI learns from data entered by humans, which can result in AI unintentionally developing biases based upon historical inequalities. For example, a company may train an AI resume scanner to identify the best candidate for a position based on previous recruiting patterns. In that case, the AI may learn that certain candidates from similar schools and life experiences are the best candidates while overlooking qualified nontraditional candidates. Legal challenges may arise when biased AI systems violate antidiscrimination laws or ethical principles. While significant efforts are being made to address the potential problem of bias with AI, for now, lawyers must be aware of the potential for biased outcomes from AI systems. There are several outstanding action items that must be addressed with respect to the integration of AI into legal practice. First, how does the use of AI in legal practice impact the ethical rules applicable to our practices, if at all? Do courts have to start regulating the use of AI in litigation, or do the existing rules appropriately account for its use? When mistakes are made, who should be held liable for the mistakes: the lawyer, the AI developer, or both? Next, does the billable hour model need to be modified in some way to account for the efficiencies we are already anticipating and expect to see in the future? Finally, there are significant confidentiality, privacy, and bias issues that need to be addressed before AI becomes more prevalent in our practice.
AI is nevertheless becoming more useful—and potentially indispensable—for legal practice. Its integration into legal practice cannot be avoided, nor should it be. The belief that AI is going to replace lawyers, particularly in litigation, is likely misguided. It is difficult to imagine that AI can replace the necessary legal judgment and discretion that is exercised by lawyers every day in furtherance of their clients’ respective goals. Rather, AI may be seen as a boon to the practice of law; it is something that can make lawyers more productive, not obsolete.
- https://www.dri.org/docs/default-source/dri-white-papers-and-reports/ai-legal-practice.pdf.
