Mexico is on the verge of enacting its first comprehensive law governing artificial intelligence — a sweeping federal framework that would classify AI violations in three tiers of severity, create new national regulatory institutions, and impose penalties ranging from fines to prison sentences. The legislation, drafted by the Senate’s specialized AI committee, represents the most ambitious attempt yet to bring legal order to a technology that has already become a vehicle for fraud, electoral manipulation, and gender-based violence.
The bill is not incremental. It establishes an entirely new category of offense — gravísimas, or “extremely serious” infractions — a term that appears in Mexican law for the first time. Acts that qualify include using AI to manipulate elections or engage in cognitive manipulation for illicit purposes, operating lethal autonomous weapons systems without human oversight, deploying mass surveillance without legal authorization, and introducing malware or backdoors into AI systems, according to Infobae. Infractions at the next level down — classified as “grave” — include running AI systems without proper certification or destroying algorithmic records.
At the same time, the bill bans creating or distributing sexually explicit deepfakes without consent. When minors are involved, no consent is possible under any circumstances. The proposal also bars AI-automated hate campaigns, defamation, smear campaigns, and gender-based digital attacks — protections that take on particular weight given that research from the Eon Institute and the Friedrich Naumann Foundation found more than 10.6 million women in Mexico suffered some form of cyberharassment between 2020 and 2025, according to Expansión.
Years in the Making
Senator Rolando Zapata, who chairs the Senate’s AI committee, calls the bill the product of 16 months of structured consultation involving 72 specialists drawn from academia, the private sector, human rights organizations, and government agencies. Every political party represented in Congress took part — Morena, PAN, PVEM, PRI, PT, and Movimiento Ciudadano all contributed to the text.
The bill proposes a national AI strategy, an AI development fund, and a federal certification system. It also sets out basic transparency requirements: because it is increasingly difficult to determine online whether content has been generated or altered using AI, the legislation would establish rules requiring that such manipulation be disclosed.
Zapata has defended the bill against accusations that it would chill free expression. “This project in no way limits or restricts freedom of expression or political debate,” he stated on his official social media account. “It does not censor opinions or criticism, nor does it establish penalties for expressing oneself.”
Significant Pushback
Critics are not convinced the language is tight enough to prevent abuse. Víctor Ruiz, CEO of the cybersecurity firm SILIKN, told Expansión that vague terms like “cognitive manipulation” and “information risks” could give authorities dangerously broad latitude. “Without precise definitions, these terms can be open to broad interpretations by the authorities,” he said.
Concerns extend beyond free speech. Mexico’s Association of the Information Technology Industry, known as AMITI, has warned that rushing AI legislation could complicate the 2026 review of the United States-Mexico-Canada Agreement, the trilateral trade pact that governs North American commerce. “We believe that any bill containing AI content might not be appropriate at this time due to the upcoming review of the USMCA,” AMITI Director General Sofía Pérez told Mexico Business News. The association, along with counterparts in the U.S. and Canada, has published a joint statement arguing that AI should be treated as an economic enabler rather than the subject of restrictive legislation — particularly before trade negotiations begin in earnest.
Where Mexico Fits Globally
The bill draws on the European Union’s AI Act, adopting a risk-based classification model that scales oversight requirements according to the potential harm a given system could cause. That framework has become a reference point for legislatures across Latin America: Brazil has advanced a similar bill through its Senate, and Chile has published AI governance guidelines, making the region a growing test case for how developing economies regulate frontier technology.
Mexico has lagged on this front despite rapid AI adoption. According to the QS World Future Skills Index 2025, AI use among Mexican companies grew by roughly 1,000% in recent years, yet around 60 AI-related bills have been introduced in Congress since 2020 without a comprehensive law emerging. The Senate bill represents the most coordinated push yet.
AI has already reached unexpected corners of Mexican life — including conservation work in Yucatán, where researchers are using the technology to track jaguars in the Dzilam de Bravo nature reserve. How the technology is governed nationally will shape its use for years across every sector.
The bill has passed its initial discussion phase and now heads to the full Senate for debate.
At a Glance: Mexico’s Proposed AI Law
- Infractions classified in 3 tiers: leves (minor), graves (serious), gravísimas (extremely serious)
- “Gravísimas” offenses include electoral manipulation, lethal autonomous systems, mass surveillance, and AI-enabled fraud
- Bans sexually explicit deepfakes without consent; no exceptions involving minors
- Prohibits AI-automated harassment, hate campaigns, and gender-based digital attacks
- Requires disclosure when AI has been used to generate or alter content
- Proposes a national AI strategy, development fund, and federal certification system
- Drafted over 16 months with input from 72 specialists; all major political parties participated
- Modeled in part on the EU’s risk-based AI Act
- Critics warn vague language could be used to restrict political speech
- AMITI warns rushed legislation could complicate 2026 USMCA trade review
- Passed initial discussion phase; pending full Senate debate
Sources: Infobae, Expansión, Mexico Business News