Lawyering in the LLMs Era: EU/US Perspectives

Large Language Models (LLMs) have emerged as a revolutionary technology in various industries, with the legal sector being one of the most promising areas of adoption. LLMs, such as OpenAI’s GPT-4 and Anthropic’s Claude, represent a distinct and advanced category of artificial intelligence (AI) due to their ability to understand and generate human-like text from vast datasets. Unlike traditional AI systems, which rely on predefined, rule-based algorithms to execute tasks, LLMs are trained on diverse and extensive linguistic data. This training enables them to generate contextually relevant, coherent, and accurate outputs, allowing for a far broader range of applications, especially in the legal profession. For example, their adaptability and fluency make them invaluable tools for automating tasks such as contract drafting, legal research, case analysis, and even preliminary legal advice. However, while LLMs offer significant potential to enhance the efficiency of legal work, their deployment in legal services also presents complex challenges in different jurisdictions. As the use of these technologies increases, it is essential to ensure their responsible integration into the legal profession, safeguarding the integrity and accountability of legal practice. The ability of LLMs to generate human-like legal text based on patterns in historical data brings both opportunities and risks. On one hand, these models can streamline workflows, reduce operational costs, and democratise access to legal resources by providing affordable legal assistance. On the other hand, the risks associated with inaccuracies, biases, and privacy breaches cannot be overlooked. These tools, if misused or inadequately supervised, could lead to errors in legal decision-making or undermine the fairness of legal proceedings. This technological shift also raises important questions for regulatory bodies and bar associations, both in the European Union (EU) and United States (US), about the scope of legal practice and the boundaries of unauthorised legal practice (ULP). As LLMs become more proficient in handling complex legal tasks, questions arise regarding the extent to which AI can be involved in providing legal services without violating the unauthorized practice of law. These questions are particularly pertinent in jurisdictions, for example the EU member states, where the legal profession is tightly regulated, with strict boundaries around who is allowed to offer legal advice and representation. Regulators must grapple with whether AI tools can ever take on roles traditionally reserved for licensed attorneys and, if so, under what conditions. Thus, this chapter argues that the increasing use of LLMs requires to reconsider the traditional regulatory models that govern legal practice, which were largely developed before such advanced AI technologies were conceived. Moreover, the regulation of LLMs for legal services introduces another layer of complexity. Both the EU and US have started developing frameworks to regulate AI technologies, but these approaches differ significantly in their scope, enforcement, and focus. In the EU, the AI Act is an effort to regulate AI technologies by categorizing systems according to their level of risk. In contrast, the US has adopted a more fragmented regulatory approach, relying largely on state-level laws. These differences in regulatory approaches create challenges for both legal professionals and technology developers in ensuring that LLMs comply with local laws and ethical standards, especially in a globalized legal environment. The responsibility for ensuring quality and fairness in LLMs-assisted legal work is an issue that cannot be addressed solely by regulation. While laws and ethical guidelines, including the rules of professional conduct of the American Bar Association (ABA) and the Council of Bars and Law Societies of Europe , provide a framework, it is ultimately up to the legal professionals to ensure that AI tools are used appropriately and ethically. Lawyers must take on an active role in ensuring that the deployment of AI in their practice is consistent with the professional duties they owe to their clients, the courts, and society at large. This shift requires that legal professionals expand their expertise to include not just traditional legal knowledge, but also an understanding of the ethical implications of using AI tools. Lawyers must learn how to effectively supervise AI-driven legal work, assess its outputs critically, and address any ethical concerns that may arise, such as unintended biases or errors in AI-generated legal advice. As the profession adapts to the increasing role of LLMs, the training and education of future lawyers will play a pivotal role. Legal education needs to evolve to integrate AI literacy, ensuring that lawyers are equipped not only with legal knowledge, but also with the skills to understand and navigate the use of AI in practice. For example, continuing legal education programs will need to focus on AI ethics, data privacy, and the responsible use of technology, helping lawyers remain informed about emerging trends and ensuring that they can exercise the ethical oversight necessary when using AI tools