API
< Back to all clusters
[TECHNOLOGY] · Germany, United States · 2 sources

started · updated

Stanford scholars warn AI could exploit legal loopholes

Felix Ringe, a visiting fellow at Stanford Law School, and co-author Linus Folkerts have warned that large language models (LLMs) are now capable of identifying exploitable inconsistencies within legal texts at scale.

In a recent paper, the researchers argued that because written law functions as the foundation of social order, it contains vulnerabilities similar to software bugs. While courts traditionally bridge these gaps through statutory interpretation, the automated discovery of legal contradictions by AI could overwhelm the judicial system's capacity to respond.

To demonstrate this, the team used LLMs to examine German administrative orders, successfully identifying several instances where regulations cited incorrect legal authorizations. The researchers suggest that legislative bodies should implement AI-driven reviews of draft regulations before they are officially published to prevent such vulnerabilities from entering the legal system.

Entities

Felix Ringe · Linus Folkerts · Stanford Law School