Thinking · AI Regulation
Argentina is building its AI regulation on EU blueprints.
Here is what the provincial patchwork means for organisations deploying AI in Latin America.
Organisations operating in Argentina face an AI regulatory landscape that is fragmented, fast-moving, and increasingly familiar. While no national AI law exists yet, four provinces have already enacted binding AI protocols — and the frameworks they chose mirror the EU AI Act almost exactly. Here is what is happening, what is likely to happen next, and what it means for governance teams planning ahead.
The state of play
Argentina does not have a single, comprehensive AI law. What it has is a layered system: a national data protection law from 2000 (Law 25,326), a set of non-binding ethical guidelines from 2023, several pending national bills, and — most significantly — a growing body of provincial regulation that is already enforceable.
Four provinces have moved first: Buenos Aires Province, the City of Buenos Aires (CABA), Santa Fe, and Jujuy. Others — Río Negro, San Juan, and San Luis — have adopted judicial AI protocols. The result is a regulatory patchwork that looks chaotic from a distance but reveals a clear directional pattern up close.
That pattern is the EU AI Act.
The EU AI Act influence is explicit
Argentina's emerging AI framework is not merely inspired by European regulation. In several cases, it directly adopts the EU AI Act's architecture.
Buenos Aires Province's Resolution 9/2025 is the clearest example. It classifies AI systems into four risk tiers — unacceptable, high, limited, and null — mirroring the EU AI Act's own taxonomy. It mandates impact assessments, requires human oversight, and prohibits AI systems deemed to pose unacceptable risk. This is not soft guidance. It is binding for the entire provincial public administration.
CABA's Resolution 206/2025 takes a similar approach for the judiciary, establishing a risk-based framework with mandatory documentation, impact assessments, and transparency requirements for any AI system used in judicial processes.
At the national level, Bill 4243-D-2025 — introduced in Congress in August 2025 — proposes a National AI Registry, mandatory risk assessments for medium- and high-risk AI systems, transparency rules, and extraterritorial application to anyone processing Argentine personal data, regardless of where the operator is located. If it passes, it will be the closest equivalent to the EU AI Act in Latin America.
Province-by-province alignment
Not every province has moved at the same pace or depth. The table below maps the key provisions of the EU AI Act against the current state of regulation in each Argentine jurisdiction.
| EU AI Act provision | Buenos Aires Province | CABA | Santa Fe | Jujuy | Río Negro | San Juan | National (pending) |
|---|---|---|---|---|---|---|---|
| Risk-based classification (Art. 5–7) | ✅ Full four-tier system | ◐ Risk-based, no formal tiers | ◐ Risk by use-case | ◔ Mentioned | ◔ Minimal | ◔ Minimal | ◷ Bill 4243-D-2025 |
| Prohibited AI practices (Art. 5) | ✅ Unacceptable risk banned | ◐ No auto-decisions on rights | ◐ No unsupervised decisions | ○ None | ○ None | ○ None | ◐ Facial recognition suspended |
| Human oversight (Art. 14) | ✅ Mandatory | ✅ Mandatory | ✅ Mandatory | ✅ Mandatory | ✅ Mandatory | ✅ Mandatory | ✅ National principle (2023) |
| Transparency & disclosure (Art. 13, 50) | ✅ Full documentation | ✅ Mandatory AI disclosure | ◐ Principles stated | ◐ General | ◐ General | ◐ Limited | ◷ National AI Registry proposed |
| Data protection (Art. 10, GDPR) | ✅ Law 25,326 aligned | ✅ Judicial data safeguards | ✅ Core to protocol | ◐ Privacy principles | ◐ National law baseline | ✅ Strict anonymisation | ◐ Reform bills pending |
| Impact assessments (Art. 9, 27) | ✅ Mandatory | ✅ Mandatory for procurement | ◐ Mentioned, not formalised | ◐ Via Innovation Lab | ○ None | ○ None | ◷ Proposed medium/high-risk |
| Registration & documentation (Art. 49, 71) | ◐ Documentation required | ◐ Procurement controls | ◔ Authority designated | ○ None | ○ None | ○ None | ◷ National AI Registry proposed |
| Regulatory authority (Art. 64–70) | ✅ Undersecretariat of Digital Gov | ◐ Council of the Judiciary | ◐ Authority of Application | ◐ Supreme Court + Inlab | ◔ Superior Tribunal | ◐ Court of Justice | ◷ National committee proposed |
| Penalties & enforcement (Art. 99–101) | ◐ Administrative discipline | ◐ Administrative discipline | ◔ General conduct rules | ◔ Judicial discipline | ◔ No AI-specific | ◔ No AI-specific | ◷ Fines and audit powers proposed |
| AI literacy & training (Art. 4) | ✅ Mandatory staff training | ✅ Judicial training | ◐ Operational rules | ✅ University collaboration | ◐ General guidance | ◐ Via JusticIA platform | ◐ ArgenIA + UNESCO |
| Extraterritorial scope (Art. 2) | ○ Provincial only | ○ CABA judiciary only | ○ Provincial only | ○ Provincial only | ○ Provincial only | ○ Provincial only | ◷ All Argentine data |
| Sector scope (Art. 2, 6) | ◐ Public admin | ◔ Judiciary only | ◐ Public admin + judiciary | ◔ Judiciary only | ◔ Judiciary only | ◔ Judiciary only | ◷ Horizontal, all sectors |
What the pattern reveals
Three things stand out.
First, human oversight is universal. Every jurisdiction — without exception — requires that AI outputs be reviewed by a human before they can inform a decision. This is the one area where Argentina is already fully aligned with the EU AI Act, from the most advanced province to the most basic judicial protocol. Organisations building AI governance infrastructure for Argentina can treat human oversight as a settled requirement.
Second, Buenos Aires Province is the benchmark. Its Resolution 9/2025 is the most comprehensive sub-national AI regulation in Latin America. It is mandatory, it covers the full public administration, and it uses the EU AI Act's own risk classification system. Any organisation deploying AI for Buenos Aires provincial government — or likely to do so — is already operating under a regulatory framework that closely mirrors the EU's.
Third, the national law will close the gaps. The pending bills — particularly Bill 4243-D-2025 — propose exactly the provisions that provincial regulations lack: a National AI Registry, extraterritorial scope, cross-sector application, and formal penalties. The direction is clear: Argentina is building toward a comprehensive, risk-based AI regulation modelled on the EU AI Act. The only question is timing.
What this means for organisations
For enterprises operating across jurisdictions, the strategic calculus is straightforward.
If your organisation deploys AI systems that process Argentine personal data — whether you are based in Buenos Aires, Amsterdam, or Singapore — you are already subject to Argentina's data protection law and will likely fall under the scope of the proposed national AI legislation. The extraterritorial clause in Bill 4243-D-2025 is modelled directly on the GDPR's approach.
If you already have EU AI Act governance infrastructure in place, most of it transfers. The risk classification system is the same. The transparency and documentation obligations are structurally similar. The human oversight requirements are identical. Building a separate governance programme for Argentina would be duplication, not compliance.
The more productive approach is to extend existing EU AI Act compliance frameworks to cover the Argentine requirements as they crystallise — and to start that work now, while the regulatory landscape is still forming and early movers can shape how governance standards are interpreted in practice.
What this means for Verdix
Verdix is built on the risk-based governance architecture that both the EU AI Act and Argentina's emerging regulatory framework share. The same structured assessment, documentation, and oversight infrastructure that meets EU requirements can be configured to address Argentine obligations — provincial and national — as they come into force.
For organisations operating in both markets, this eliminates the need for parallel governance systems. One platform. One evidence base. Multiple regulatory frameworks.
This article is based on publicly available legislative texts, provincial resolutions, and regulatory analysis current as of June 2026. Argentina's national AI legislation remains under congressional deliberation. Provincial regulations apply within their respective jurisdictions and sectors. This article is for informational purposes only and does not constitute legal advice. Organisations should consult qualified legal counsel on the specific obligations applicable to their AI systems in Argentina.