Harry Gimson (Swansea University – Swansea University) has posted Legal Reasoning in the Age of LLMs: Progress, Problems, and Professional Implication on SSRN. Here is the abstract:
This essay compresses the trajectory from early symbolic legal AI to large language models (LLMs), critically examining technical designs, inferential limits, practitioner uptake, and policy implications. It argues that while LLMs solve many scale and flexibility problems, they do not eliminate deeper representational and normative deficits in legal reasoning.