Mexico - AI Legislation Empowerment (2025)

Overview

The initiative "Que reforma la fracción XVII del artículo 73 de la Constitución Política de los Estados Unidos Mexicanos, en materia de inteligencia artificial" was submitted by Deputy Ricardo Monreal Ávila and published in the Gaceta Parlamentaria on 19 February 2025. It proposes to modify Article 73 to explicitly add "intelligence artificial" (artificial intelligence) to the list of matters on which the Congress of the Union may legislate, thereby providing a constitutional basis for a future General Law regulating AI and its uses across public and private sectors. The initiative stresses that an express constitutional grant would reduce constitutional litigation risk, enable coherent national policy, and facilitate alignment with international frameworks (UN, EU, OECD). The draft includes two transitional provisions proposing short deadlines for harmonizing secondary legislation and for issuing a General Law on the Use of Artificial Intelligence after the decree enters into force. The published text and exposition of motives are available in the Chamber of Deputies' Gaceta. ([gaceta.diputados.gob.mx](https://gaceta.diputados.gob.mx/Gaceta/66/2025/feb/20250219-II-1.html))

Definitions

Key terms used in the initiative are drawn from the exposition of motives and legislative drafting practice. "Artificial intelligence (AI)" is referenced as the discipline and set of systems enabling machines to perform tasks that would normally require human intelligence (learning, reasoning, natural language processing). "General Law" refers to a future federal statute (Ley General en materia de Uso de Inteligencia Artificial) envisaged by the transitory provisions. "Harmonization" refers to the process by which Congress and relevant authorities adapt or amend secondary laws to align with constitutional change. The initiative does not supply exhaustive technical definitions (it leaves definitional detail to the future General Law), but it frames AI as a technology with cross-sectoral impacts requiring ethical, privacy and security-oriented regulatory treatment. ([gaceta.diputados.gob.mx](https://gaceta.diputados.gob.mx/Gaceta/66/2025/feb/20250219-II-1.html))

Governance and Institutional Framework

The reform is procedural and constitutional in nature: it empowers the Congress of the Union to legislate on AI, but implementation will require new or adapted institutional roles. The initiative explicitly contemplates coordination with the executive branch bodies named in the exposition—particularly the newly established Agencia de Transformación Digital y Telecomunicaciones (ATDT) and the Secretaría de Ciencia, Humanidades, Tecnología e Innovación (SECIHTI)—which the exposition identifies as key interlocutors for evidence-based lawmaking and policy development. Once the constitutional amendment is approved and the General Law enacted, governance functions should include: (i) a federal coordinating mechanism for AI policy; (ii) technical standard-setting and conformity assessment units (public or delegated to recognized bodies); (iii) supervisory and market surveillance authorities; (iv) data protection and privacy oversight alignment with the Federal Law on Protection of Personal Data (and its authorities); and (v) inter-agency mechanisms for critical infrastructure, national security, and ethical review. The Gaceta emphasizes dialogue with scientific community and sector stakeholders as part of developing the General Law. ([gaceta.diputados.gob.mx](https://gaceta.diputados.gob.mx/Gaceta/66/2025/feb/20250219-II-1.html))

Key Focus Areas

The initiative identifies several thematic areas a future General Law should address: (1) respect for fundamental rights (non-discrimination, due process, freedom of expression, dignity); (2) data protection and privacy safeguards; (3) transparency and explainability requirements for high-impact systems; (4) risk-based classification of AI systems to prioritize regulatory attention on high- and critical-risk use cases (e.g., biometric ID, public safety, healthcare diagnostics, employment and credit scoring); (5) human oversight and meaningful control over automated decisions; (6) mandatory documentation, auditing and record-keeping (model cards, datasets provenance, training logs) to ensure accountability and reproducibility; (7) conformity assessment and registration requirements for certain categories of systems; (8) security and resilience measures addressing model theft, poisoning, and adversarial attacks; (9) consumer protection, liability and remedies for harm caused by AI systems; (10) public procurement and government use policies, including safeguards for public-interest deployments; and (11) research, innovation and the promotion of technological sovereignty (support for domestic R&D and responsible data-sharing frameworks). These focus areas mirror international trends and are referenced in the initiative's comparative analysis of EU, U.S., and multilateral approaches. ([gaceta.diputados.gob.mx](https://gaceta.diputados.gob.mx/Gaceta/66/2025/feb/20250219-II-1.html))

Implementation Framework

The initiative's transitory articles set a short implementation timeline: entry into force the day after publication; 90 days to harmonize secondary legislation (in one transitory clause as published); and 180 days to promulgate a General Law on the Use of Artificial Intelligence (in a second transitory clause). Operationalizing these deadlines will require a rapid legislative drafting process, stakeholder consultations, and likely delegation to specialized units within the ATDT and SECIHTI for technical standards and technical assistance. The legislative process should adopt a risk-based regulatory design: define categories of AI systems, determine obligations by category, set conformity assessment procedures, and assign supervisory roles. The General Law should also create mechanisms for regulatory sandboxes, prioritized public investments in data infrastructure, and guidance for public sector use. The short statutory timelines in the initiative mean Congress would need to coordinate across committees and with the executive to produce technically sound secondary legislation without undermining procedural safeguards. ([gaceta.diputados.gob.mx](https://gaceta.diputados.gob.mx/Gaceta/66/2025/feb/20250219-II-1.html))

Monitoring and Evaluation

The initiative underscores the need for ongoing monitoring of AI deployments and periodic evaluation of the legal framework. Monitoring mechanisms should include mandatory incident reporting for serious harms, an AI system registry for high-risk applications, periodic impact assessments for regulated categories, and requirements for independent audits. Evaluation metrics should address: rights-protection outcomes; incidence of algorithmic discrimination; cybersecurity incidents affecting AI systems; compliance rates with transparency and documentation obligations; and the effectiveness of sanctions. The transitory text does not establish a specific supervisory agency or metrics, leaving these design choices to the General Law and delegated authorities. The Gaceta recommends consultation with the scientific and civil society communities during the design of monitoring frameworks. ([gaceta.diputados.gob.mx](https://gaceta.diputados.gob.mx/Gaceta/66/2025/feb/20250219-II-1.html))

Penalties, Liability, and Appeals

The initiative itself does not specify granular sanctions but anticipates that the General Law will provide penalties and liability rules proportionate to harm and culpability. Expected instruments include administrative fines, suspension or removal of offending AI systems from the market, revocation of registrations or certifications, compensatory remedies for harmed individuals, and criminal sanctions for intentional, malicious, or fraudulent conduct. An appeal mechanism within the administrative and judicial system is necessary to safeguard due process: regulated entities should have access to administrative review and judicial protection (amparo). The General Law should also clarify allocation of civil liability between AI developers, deployers/operators, and data providers, including rules on strict vs fault-based liability in high-risk contexts. ([gaceta.diputados.gob.mx](https://gaceta.diputados.gob.mx/Gaceta/66/2025/feb/20250219-II-1.html))

Relationship to Other Instruments

The proposed constitutional change is expressly designed to enable a federal General Law on AI and to reduce legal uncertainty when harmonizing existing statutes (telecommunications, data protection, consumer protection, public procurement, health, labor, criminal law). The initiative highlights the need to coordinate with the Federal Law on Protection of Personal Data Held by Private Parties, the Ley General de Salud, the Ley de Obras Públicas (where relevant) and penal provisions addressing misuse. The transitory provisions call for harmonization of secondary law, which will require careful review to avoid overlaps and to assign clear competencies between the Congress, ATDT, SECIHTI, sectoral regulators (e.g., financial regulator for algorithmic credit decisions), and newly restructured telecommunications regulators. ([gaceta.diputados.gob.mx](https://gaceta.diputados.gob.mx/Gaceta/66/2025/feb/20250219-II-1.html))

International Alignment

The initiative explicitly cites international developments—UN resolution on AI, the EU's risk-based regulatory model, and U.S. policy initiatives (AI Bill of Rights and federal/state actions)—as benchmarks and argues that an explicit constitutional competence will help Mexico align with international norms and obligations. The drafting approach recommended in the exposition advocates for a risk-based, rights-respecting, and innovation-friendly law designed to promote interoperability with global standards and facilitate cross-border data flows while preserving privacy and security. International cooperation (standard-setting bodies, multilateral fora) and mutual recognition for conformity assessments are anticipated policy instruments for alignment. ([gaceta.diputados.gob.mx](https://gaceta.diputados.gob.mx/Gaceta/66/2025/feb/20250219-II-1.html))

Implementation Timeline

Milestone	Date / Deadline
Initiative published in Gaceta Parlamentaria (Chamber of Deputies)	2025-02-19
Proposed transitory entry into force	Day after publication in the Diario Oficial de la Federación (if approved)
Congress to harmonize secondary laws (transitory)	Within 90 calendar days after entry into force
Congress to promulgate General Law on the Use of AI (transitory)	Within 180 calendar days after entry into force

Compliance Checklist

Requirement	Notes
Determine whether system is high-risk	Classify by sector and impact on rights
Conduct and publish risk/impact assessment	Data protection impact assessment + AI-specific risk analysis
Register high-risk systems (if required)	National registry or supervisory database
Implement human oversight measures	Explainable outputs and human-in-the-loop controls
Maintain technical documentation	Model cards, training data summaries, audit logs

Sources and References

Source	Type
Gaceta Parlamentaria: Iniciativa para reformar la fracción XVII del artículo 73 (Ricardo Monreal Ávila), 19 Feb 2025	Primary Source
Gaceta Parlamentaria: Iniciativa (PAN) reforma Art. 73 en materia de IA, 22 Apr 2025	Primary Source
Agencia de Transformación Digital y Telecomunicaciones (official site)	Primary Source
DOF / Agreement: Domicilio oficial and functions of the ATDT (Diario Oficial references)	Primary Source