Buenos Aires court drafts rules for responsible judicial use of AI

Argentina’s largest provincial court is doing what many systems talk about but few implement: writing concrete house rules for AI in day‑to‑day judicial work. The piece describes a reality where lawyers already use ChatGPT for pleadings, experts draft opinions with AI assistance and clerks lean on LLMs for jurisprudence searches, all in a regulatory vacuum. The SCBA’s draft tries to close that gap by clarifying when AI can be used, what transparency is required and where human responsibility cannot be delegated.

This matters to the AGI conversation because courts are among the most sensitive environments for deploying advanced models. If a provincial supreme court normalizes AI‑assisted drafting but insists on traceability and human accountability, it could become a reference for other civil‑law jurisdictions grappling with similar pressures. It also pushes back on both extremes — outright bans and uncritical adoption — by treating AI as a tool whose risks can be managed procedurally.

For global labs, it’s a signal that “vertical” governance of AI use cases is arriving, one professional community at a time. Future, more capable models will inherit these norms, shaping how far automated legal reasoning is allowed to go before courts insist on a human in the loop.