Artificial Intelligence and Law – An Overview of Recent Technological Changes in Large Language Models and Law

Abstract

This article, based upon a keynote address by Professor Harry Surden, provides an in-depth overview of the recent advancements in artificial intelligence (AI) since 2022, particularly the rise of highly capable large language models (LLMs) such as OpenAI’s GPT-4, and their implications for the field of law. The talk begins with a historical perspective, tracing the evolution of AI from early symbolic systems to the modern deep learning era, highlighting the breakthroughs that have enabled AI to process and generate human language with unprecedented sophistication.

The address explores the technical foundations of contemporary AI models, including the transition from rule-based systems to data-driven machine learning, the role of deep learning, and the emergence of transformers, which have significantly enhanced AI’s ability to understand and generate text. It then examines the capabilities and limitations of GPT-4, emphasizing its strengths in legal research, document drafting, and analysis while also identifying key concerns, such as hallucinations, biases, and the sensitivity of AI outputs to user prompts.

The article also considers the potential risks of using AI in legal decision-making, particularly in judicial settings, where AI-generated legal reasoning may appear authoritative yet embed subtle interpretive choices. It argues that while AI can assist in legal tasks, it should not be treated as a neutral arbiter of law. It concludes by addressing near-term trends in AI, including improvements in model accuracy, interpretability, and integration into legal workflows, and emphasizes the need for AI literacy among legal professionals.

Keywords: Artificial intelligence, ChatGPT, LLMs, Large Language Models, AI, Automation, AI and Law