Algorithmic Challenges and Regulatory Frameworks of Artificial Intelligence in Mexico: A Prospective Analysis from the Perspective of Digital Governance Theory

Abstract

The rapid integration of artificial intelligence (AI) has heightened the need for evidence-based regulatory frameworks to effectively address its legal, ethical, and societal consequences. This research carefully analyzes the prevailing landscape of AI-related legislation in Mexico. The study conducts a comprehensive review of legislative initiatives related to AI regulation submitted to Mexican legislative bodies, encompassing those approved or pending in commissions. This process leads to the identification and categorization of outstanding initiatives across seven policy areas: Congress, Education, Health, Intellectual Property, Justice, AI Promotion, and AI Regulation. As a principal contribution, this work offers the first exhaustive mapping and thematic classification of legislative activity related to AI in Mexico. Furthermore, the analysis identifies systemic regulatory deficiencies, such as the lack of AI-specific legislation, the limited scope of existing data protection laws in relation to AI systems, and an absence of technical provisions concerning ethical design, algorithmic transparency, cybersecurity, and accountability frameworks. By showcasing these deficiencies, the study contributes a diagnostic framework for evaluating AI governance readiness in emerging economies. The findings emphasize the importance of establishing a comprehensive, technically sound, and internationally harmonized regulatory framework to reduce AI-related risks while promoting responsible innovation in Mexico.

1. Introduction

Artificial intelligence (AI) has become an integral part of modern life. Recent advances in generative transformers, along with other technologies, are reshaping society, driving rapid development across various domains. However, these advances do not come without limitations and concerns regarding their safety, environmental impact, and potential disruptive influence on jobs, the economy, and other social aspects [1].

In addition, several concerns have been raised about autonomous weapons, including the potential for chemical and biological attacks, AI-based cyberattacks, and the facilitation of organized crime by AI, as well as the ultimate accountability of technologies that incorporate AI [1] and its ethical use [2]. That is why the rapid advancement of AI technologies has sparked a global dialogue on the need for comprehensive legal frameworks to regulate their use, and Mexico is no exception. As AI systems continue to permeate diverse sectors, from healthcare to finance, the Mexican legal landscape faces a pressing need to adapt and evolve to address the complexities and ethical dilemmas posed by these technologies.

Mexico was among the first nations to initiate discussions on AI regulation in 2018, emphasizing the need for each country to define its own AI regulatory framework, reflecting a proactive stance in shaping ethical guidelines tailored to its national priorities [3]. The concerns regarding digital governance in Mexico have been studied for several years. A pioneering study by Schneider concludes that, although Mexico has established an AI and digital transformation strategy, its concrete agenda remains in its early stages. The country lacks a large market to exert significant pressure on major digital platforms, and large corporations can easily bypass existing Mexican data protection laws (Federal Law on the Protection of Personal Data Held by Private Parties [4]). In addition, enforcing the current regulations remains an area for improvement [5]. Ruvalcaba-Gomez and Cifuentes-Faura studied the perception of digital governance and artificial intelligence in the state of Jalisco, Mexico [6]. Their study showed that public officials consider themselves fairly proficient in information technologies but less proficient in AI-related issues. It was reported that 53.8% of respondents have never heard of AI in the public sector, and 98.9% consider it necessary to develop regulations and public policies on AI.

Other researchers studied the impact and barriers of AI in the public sector, as well as the perception of local politicians [7,8]. They analyzed public officials’ perceptions of the impact, functions, and barriers of artificial intelligence (AI) within the context of a local government in the State of Mexico. A survey of 32 key public managers identified the competencies, perceptions, and challenges of AI implementation. These studies reveal concerns, functional impairments, and key obstacles to AI implementation. While AI has the potential to significantly transform and improve the efficiency and quality of public services, it also faces significant barriers that must be addressed to ensure its success. Appropriate regulation and ongoing training for public officials are crucial to fully harness the benefits of AI and mitigate its potential adverse effects.

In addition, Criado et al. conducted a survey administered to chief information officers in ministries of Mexico and Spain [9]. The survey identified three key inhibitors to the adoption of AI in the Mexican public sector: budget (90.9%), the digital divide (63.6%), and technological infrastructure (54.5%). An intriguing finding is that security and data privacy were not perceived as barriers to AI implementation in the public sector, as none of the Mexican respondents selected these options. Similar concerns regarding the adoption of AI in the public sector of different countries from an ethical perspective were summarized in 2021 as part of the CLAD International Course “Governance of Artificial Intelligence in Public Administration from an Ethical Perspective” [10].

An interesting study explored users’ perceptions of AI-based chatbots during the COVID-19 pandemic [11]. They conducted an open survey regarding two chatbots (Susana Distancia and Dr. Armando Vaccuno) deployed as part of a “digital government” strategy of Mexican health systems. The survey reported high levels of apprehension regarding the transparency, data protection, and accountability of the chatbots. Both users and non-users share concerns about the lack of information regarding the treatment and processing of the collected data, the lack of credibility in governmental responsibility for the bots’ functioning, and a high level of concern about the use, distribution, and protection of their personal data.

Additionally, Mendoza-Enríquez commented on the challenges related to data protection. She considered that, although Mexico has a solid legal framework for personal data protection, it faces challenges in technological environments, particularly with the use of AI. The extraterritoriality of the law, the absence of physical borders, and the reliance on corporations to safeguard rights diminish its effectiveness. Furthermore, new forms of the right to data protection need to be recognized, such as the right to challenge automated decisions, the right not to be subjected to AI processing, and the right to non-identification in these processes [12]. A small survey by Pérez et al. in Yucatan on AI safety revealed concerns about the spread of misinformation, manipulation of public opinion, and the possibility of AI being used in an authoritarian manner [13].

Despite those previous investigations, there is no comprehensive analysis of the current legislation governing artificial intelligence in Mexico. In this context, questions arise regarding the current status of different regulatory proposals, the adequacy of current regulations, and the effectiveness of enforcement mechanisms. Furthermore, significant challenges remain in implementing AI regulations, including striking a balance among innovation, ethical considerations, and the protection of individual rights.

This review paper will examine the current legal framework governing AI in Mexico by reviewing the existing initiatives regarding AI regulation and discussing the potential future directions for the regulatory landscape. The aim of this review is to analyze the current state of legislative initiatives in Mexico related to the development, use, and regulation of artificial intelligence.

We want to focus on two research questions:

Q1: 

What legislative initiatives currently exist in Mexico at the federal level that directly or indirectly address the use of artificial intelligence?

Q2: 

What critical aspects of AI (such as algorithmic bias, transparency, accountability, or privacy) are insufficiently regulated or completely absent in current legislation?

By reviewing the current state of the AI-related legislative corpus in Mexico and addressing the research questions above, our study will ultimately contribute to a more informed and nuanced understanding of the country’s AI governance.

The rest of the paper is organized as follows: Section 2 assesses the legal framework governing AI in Mexico till 15 May 2025. Section 3 discusses the initiatives reviewed and the limitations of our study. Section 4 summarizes key challenges and future directions in AI regulation in the studied country and presents actionable recommendations. Finally, Section 5 concludes the paper.

2. Assessment of the Legal Framework Governing AI in Mexico

The QS World Future Skills Index 2025 report stated that Mexico increased its use of AI recently, with over 40% of Mexican companies already using AI, and job offers in the field soared by 95% in 2023 [14]. Although Mexico unveiled its AI strategy as early as 2018, the country has yet to enact specific laws governing AI, despite growing legislative interest in this area.

Artificial intelligence offers real opportunities for improvement in multiple areas, but using it without clear rules can affect rights, freedoms, and social trust. Regulating it does not mean stifling innovation; instead, it means establishing reasonable limits to prevent abuse, fraud, manipulation, or harm to individuals.

From an ethical perspective, technology must serve human dignity. Decisions relevant to people’s lives—such as those related to employment, health, education, finances, or security—cannot be left entirely to automated systems without human oversight, transparency, and accountability [15].

A risk-based regulatory model allows stricter controls to be applied only where they are truly needed, avoiding unnecessary burdens on low-impact uses. In high-risk cases, prior assessment, explanation of decisions, identification of AI-generated content, and the possibility of holding those who design or use it accountable are essential.

In this regard, as stated by the European Union Artificial Intelligence Act, a general law for AI can “improve the functioning of the internal market and promote the uptake of human-centric and trustworthy artificial intelligence (AI), while ensuring a high level of protection of health, safety, fundamental rights (…), including democracy, the rule of law and environmental protection, against the harmful effects of AI systems (…) and supporting innovation” [16]. Its purpose is not to control people, but to protect them, ensuring that artificial intelligence contributes to social well-being, strengthens public trust, and respects the fundamental principles of the rule of law.

Reviewing the initiatives currently available for analysis in the legislative corps is urgent and necessary for several reasons:

• Keeping up with technological advancement: Current laws were not designed to address the unique challenges posed by contemporary AI, such as machine learning, automated decision-making, or the use of generative models.
• Protection of human rights: AI can reinforce structural inequalities if not properly regulated. It is vital to guarantee the protection of privacy, non-discrimination, due process, and other fundamental rights against the misuse of automated systems.
• Guidance from international laws: Mexico needs to harmonize its legislation with global frameworks to foster international trust, facilitate cross-border cooperation, and actively participate in multilateral discussions on AI.
• Promoting ethical and responsible innovation: A clear regulatory framework can provide legal certainty for developers and companies, promoting technological innovation based on ethical principles and with social responsibility.
• Prevention of risks and harm: AI applications in security, surveillance, education, or justice can have significant consequences if not subject to appropriate controls. Legislative review must allow for effective mechanisms for impact assessment, oversight, and accountability.

Numerous initiatives focused on artificial intelligence have been proposed, some of which are monitored by the National Alliance for artificial intelligence [17]. Additionally, the Legislative Information System (Sistema de Información Legislativa, SIL, in Spanish) website provides information on the current status of all initiatives presented in various legislative scenarios at the federal level.

For our review, we have the following inclusion and exclusion criteria:

Inclusion Criterion: Initiatives introduced in a Mexican legislative body that refer to the use of artificial intelligence.

Exclusion Criteria:

• Initiatives rejected in committees.
• Initiatives withdrawn from the legislation.

The Legislative Information System is a website hosted by the federal government, and we selected it as the primary data source for the review because of the following:

• It contains all initiatives presented in all legislative scenarios.
• It is an official website, hosted by the Secretary of the Interior (Secretaría de Gobernación, in Spanish).
• It contains a search engine that is able to query the information by using text.

As a search strategy (Figure 1), we queried the SIL website on 21 May 2025, for all legislative periods (http://sil.gobernacion.gob.mx/Busquedas/Basica/; accessed on 21 May 2025) using the words “inteligencia artificial” (artificial intelligence), and selecting “iniciativa” (initiative) as “tipo de asunto” (type of matter).

The first initiatives were proposed during the LXIV legislative period (September 2018–August 2021), following which there was an increase in interest in regulating AI. Fifty-five initiatives were presented in the LXV legislative session from September 2021 to August 2024, and forty-one initiatives have been presented in the LXVI legislative session from September 2024 to 15 May 2025.

In total, the website returned 105 initiatives. Of those, 56 had the status of “desechado” (rejected) in the corresponding committee, and one had the status of “retirado” (withdrawn); these were excluded from this review. In addition, two initiatives lacked the corresponding .pdf document in the system and could not be retrieved for full-text analysis.

Then, two independent reviewers manually screened the titles and summaries of the remaining 46 initiatives and selected 40 which were relevant. Six were discarded because, although their text mentioned the words “artificial intelligence”, they were related to other topics, as follows:

• Adding regulation to the federal law of work, related to the workplace rights of workers of digital platforms such as Uber, Amazon, and others.
• Adding regulations to the law of rural sustainable development.
• Adding regulations to the General Law for the Care and Protection of People with Autism Spectrum Disorders.
• To promote the use of information technologies that enable distance learning in situations of violence, health emergencies, or natural disasters.
• To amend and add various provisions to the General Law for the Care and Protection of Persons with Autism Spectrum Disorder.
• To stipulate that the Public Prosecutor’s Office may receive complaints or charges in person and through artificial intelligence platforms or by any other technological means.

Both reviewers agree on the 40 relevant initiatives and the six excluded ones. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram [18] of our review is provided in Figure 2. It is important to note that we only reported on 38 of the 40 studies included, because we realized that one study had a different number and URL, but the exact text in the system, and another was presented twice (same text, two presentations).

Regarding the status of the relevant initiatives, none have been approved to date (all have a “pending” status). We consider our study to be free from bias because our primary data source is the official governmental website, and it compiles all initiatives presented by all political parties in all legislative bodies (Figure 3). The political parties with the greatest level of involvement in AI regulation from 2018 to 2025 are MORENA and PAN.

The legislative information system allows users to view, for each initiative, the name (title), the object (summary), the observations, the presentation date, and the full text in PDF. We did not use any automated tools for the review, and the system provides a summary (object) for each item, making it easy to determine the item’s objective and topic. Each initiative was reviewed by two independent individuals for analysis and categorization.

Our first research question, Q1, was as follows: What legislative initiatives currently exist in Mexico that directly or indirectly address the use of artificial intelligence? To answer this, the relevant initiatives were reviewed and categorized into eight areas based on their summaries (Figure 4).

All initiatives related to criminal prosecution, crime classification, violence prevention, and prosecution in general were categorized as Justice-related. The ones aimed at regulating AI in some form were categorized as Regulate IA; the ones related to education at all levels, including personnel training, were classified as Education, the ones related to human health and using AI in health environments were categorized as Health; the initiatives related to generative art, use of the image and voice of actors, and similar topics were classified as related to Intellectual Property; the ones devoted to promote the use of AI were categorized as Promote AI. There were also other initiatives aimed at modifying the Congress’s faculties to legislate on AI, classified as Congress. The general-purpose initiatives were categorized under the Regulate AI initiative.

As shown in Figure 4, the majority of the initiatives (37%) are focused on justice-related topics, including but not limited to the misuse of generative AI for identity theft, sexual harassment, explicit content involving minors, and similar issues. Below, we analyze the initiatives by area of interest, in ascending order of the number of initiatives.

2.1. Congress-Related Initiatives

A special mention is made of initiatives to empower the Mexican Congress to enact AI-related laws. The current text of the constitutional Article 73 states that “Congress has the power: (…) XVII. To enact laws on general means of communication, information and communication technologies, broadcasting, telecommunications, including broadband and the Internet, postal services and mail, and on the use and exploitation of waters under federal jurisdiction (…)”. However, it does not explicitly include AI, just the more general terms of “information and communication technologies, broadcasting, telecommunications, including broadband and the Internet”.

Our review found nine initiatives (Table 1) from four political parties (PAN, MORENA, PT, and PRI). All were presented between 6 February and 30 April 2025. Of those, seven have very similar proposed texts, but none have gained sufficient consensus to be approved or modified; all remain pending with the commissions.

Only the first proposed initiative [19] takes a more holistic approach: it suggests a reform of articles 1, 6 and 73, including a prevention of AI discrimination (Article 1), the state obligation to promote cybersecurity and to establish a National Agency for artificial intelligence and Cybersecurity (Article 6), and the empowerment of Congress to enact laws on AI, cyber and neuro technologies, and similar (Article 73). Another difference is the proposal by Herrera-Ruiz, which, instead of modifying item XVII, adds a new fraction XXXII [20].

Table 1. Initiatives related to empowering Congress to enact laws regarding artificial intelligence (order: recent to older). The proposed text additions are presented in bold.

Ref.	Party, Date	Proposed Text
[21]	PT, 30 April 2025	Article 73. Congress has the power to: (…) XVII. To enact laws on general means of communication, artificial intelligence, information and communications technologies, broadcasting, telecommunications, including broadband and the Internet, postal services and mail, and on the use and exploitation of waters under federal jurisdiction (…)
[20]	Morena, 30 April 2025	Article 73. Congress has the power to: (…) XXXII. To enact general laws regarding emerging, disruptive technologies, and artificial intelligence, establishing the foundations, principles, and guidelines for their ethical, safe, transparent, sustainable, and human rights-respecting use, as well as to promote their development for the benefit of the nation.
[22]	PAN, 29 April 2025	Article 73. Congress has the power: (…) XVII. To enact laws on general means of communication, information and communication technologies, broadcasting, telecommunications, including broadband and the Internet, artificial intelligence, postal services and mail, and on the use and exploitation of waters under federal jurisdiction (…)
[23]	Morena, 2 April 2025	Article 73. Congress has the power: (…) XVII. To enact laws on general means of communication, information and communication technologies, broadcasting, telecommunications, including broadband and the Internet, postal services and mail, the general law on artificial intelligence that establishes principles, bases, limits and prohibitions on the use of such technology, and on the use and exploitation of waters under federal jurisdiction (…)
[24]	Morena, 20 March	Article 73. Congress has the power (…) XVII. To enact laws on general means of communication, information and communication technologies, artificial intelligence, broadcasting, telecommunications, including broadband and the Internet, mail, and on the use and use of postal services and mail, and on the use and exploitation of waters under federal jurisdiction (…)
[25]	PAN, 20 March 2025	Article 73. Congress has the power (…) XVII. To enact laws on general means of communication, information and communication technologies, broadcasting, telecommunications, including broadband and the Internet, postal services and mail, artificial intelligence, and on the use and exploitation of waters under federal jurisdiction (…)
[26]	PRI, 4 March 2025	Article 73. Congress has the power: (…) XVII. To enact laws on general means of communication, information and communication technologies, broadcasting, telecommunications, including broadband and the Internet, postal services and mail, on artificial intelligence, providing for limits and restrictions on its use, and on the use and exploitation of waters under federal jurisdiction (…)
[27]	Morena, 19 February 2025	Article 73. Congress has the power: (…) XVII. To enact laws on general means of communication, information and communication technologies, artificial intelligence, broadcasting, telecommunications, including broadband and the Internet, postal services and mail, and on the use and exploitation of waters under federal jurisdiction (…)
[19]	PAN, 6 February 2025	Article 1. (…) Any discrimination motivated by ethnic or national origin, gender, age, disabilities, social status, health conditions, religion, opinions, sexual preferences, marital status, those derived from the use of applications or tools of artificial intelligence, cybertechnology or neurotechnology, or any other that violates human dignity and aims to annul or undermine the rights and freedoms of people is prohibited. Article 6. The State shall promote cybersecurity, as well as the development and implementation of artificial intelligence technologies, tools, and applications, cybertechnologies, among other similar technologies, and shall regulate that these are safe, responsible, and protect people’s human rights, avoiding discriminatory biases and unacceptable risks, as well as balancing innovation with ethical and security standards. National law shall protect people’s neuro-rights, especially safeguarding brain activity and the information derived from it, and shall establish a National Agency for Artificial Intelligence and Cybersecurity. Article 73. Congress has the power: (…) XVII. To issue: (a) Laws regulating general means of communication, information, and communication technologies, broadcasting, telecommunications, including broadband and the Internet, postal services, and mail; (b) National laws on cybersecurity and on technologies, tools, and applications of artificial intelligence, cybertechnologies, neurotechnologies, among other similar technologies, as well as those that protect neuro-rights, in terms of the provisions of articles 1 and 6 of this Constitution. (c) The laws on the use and exploitation of waters under federal jurisdiction (…)

Table 1. Initiatives related to empowering Congress to enact laws regarding artificial intelligence (order: recent to older). The proposed text additions are presented in bold.

Ref.	Party, Date	Proposed Text
[21]	PT, 30 April 2025	Article 73. Congress has the power to: (…) XVII. To enact laws on general means of communication, artificial intelligence, information and communications technologies, broadcasting, telecommunications, including broadband and the Internet, postal services and mail, and on the use and exploitation of waters under federal jurisdiction (…)
[20]	Morena, 30 April 2025	Article 73. Congress has the power to: (…) XXXII. To enact general laws regarding emerging, disruptive technologies, and artificial intelligence, establishing the foundations, principles, and guidelines for their ethical, safe, transparent, sustainable, and human rights-respecting use, as well as to promote their development for the benefit of the nation.
[22]	PAN, 29 April 2025	Article 73. Congress has the power: (…) XVII. To enact laws on general means of communication, information and communication technologies, broadcasting, telecommunications, including broadband and the Internet, artificial intelligence, postal services and mail, and on the use and exploitation of waters under federal jurisdiction (…)
[23]	Morena, 2 April 2025	Article 73. Congress has the power: (…) XVII. To enact laws on general means of communication, information and communication technologies, broadcasting, telecommunications, including broadband and the Internet, postal services and mail, the general law on artificial intelligence that establishes principles, bases, limits and prohibitions on the use of such technology, and on the use and exploitation of waters under federal jurisdiction (…)
[24]	Morena, 20 March	Article 73. Congress has the power (…) XVII. To enact laws on general means of communication, information and communication technologies, artificial intelligence, broadcasting, telecommunications, including broadband and the Internet, mail, and on the use and use of postal services and mail, and on the use and exploitation of waters under federal jurisdiction (…)
[25]	PAN, 20 March 2025	Article 73. Congress has the power (…) XVII. To enact laws on general means of communication, information and communication technologies, broadcasting, telecommunications, including broadband and the Internet, postal services and mail, artificial intelligence, and on the use and exploitation of waters under federal jurisdiction (…)
[26]	PRI, 4 March 2025	Article 73. Congress has the power: (…) XVII. To enact laws on general means of communication, information and communication technologies, broadcasting, telecommunications, including broadband and the Internet, postal services and mail, on artificial intelligence, providing for limits and restrictions on its use, and on the use and exploitation of waters under federal jurisdiction (…)
[27]	Morena, 19 February 2025	Article 73. Congress has the power: (…) XVII. To enact laws on general means of communication, information and communication technologies, artificial intelligence, broadcasting, telecommunications, including broadband and the Internet, postal services and mail, and on the use and exploitation of waters under federal jurisdiction (…)
[19]	PAN, 6 February 2025	Article 1. (…) Any discrimination motivated by ethnic or national origin, gender, age, disabilities, social status, health conditions, religion, opinions, sexual preferences, marital status, those derived from the use of applications or tools of artificial intelligence, cybertechnology or neurotechnology, or any other that violates human dignity and aims to annul or undermine the rights and freedoms of people is prohibited. Article 6. The State shall promote cybersecurity, as well as the development and implementation of artificial intelligence technologies, tools, and applications, cybertechnologies, among other similar technologies, and shall regulate that these are safe, responsible, and protect people’s human rights, avoiding discriminatory biases and unacceptable risks, as well as balancing innovation with ethical and security standards. National law shall protect people’s neuro-rights, especially safeguarding brain activity and the information derived from it, and shall establish a National Agency for Artificial Intelligence and Cybersecurity. Article 73. Congress has the power: (…) XVII. To issue: (a) Laws regulating general means of communication, information, and communication technologies, broadcasting, telecommunications, including broadband and the Internet, postal services, and mail; (b) National laws on cybersecurity and on technologies, tools, and applications of artificial intelligence, cybertechnologies, neurotechnologies, among other similar technologies, as well as those that protect neuro-rights, in terms of the provisions of articles 1 and 6 of this Constitution. (c) The laws on the use and exploitation of waters under federal jurisdiction (…)

2.2. Education-Related Initiatives

Four initiatives were presented regarding education (including continuous capacitation).

The first was a proposal by the PAN political party that amends Article 3 of the Political Constitution of the United Mexican States [28]. It proposes to establish in the Constitution that education should be “(…) and technologically advanced”.

In addition, it adds a new item, stating that

“(d) As a means of guaranteeing the right to education, the State shall ensure that the education it provides is effectively supported through all available technological and technical tools, including information and communication technologies, artificial intelligence, educational blockchain, and high-quality broadband connectivity. To this end, schools must have the necessary technological infrastructure, always ensuring that the educational processes maintain full curricular and academic validity, as established by law, to promote equal opportunities, universal access to knowledge, and the development of a skilled and specialized workforce.”

The second initiative by MORENA political party amends Articles 153-C and 153-K of the Federal Labor Law [29] and aims to provide workers with training in skills related to artificial intelligence. It adds a new item to 153-C:

“Provide training to employees on skills related to artificial intelligence, so that they can stay up to date and learn about the use and application of artificial intelligence in businesses, thus increasing productivity.”

It also empowers the National Committee on Competitiveness and Productivity to formulate training plans and programs to:

“Develop recommendations for training and development programs that will enhance worker productivity, job skills, and overall competencies. These recommendations should also consider the use and knowledge of skills related to artificial intelligence.”

The third one (also by MORENA political party) aims to amend Article 30 of the General Education Law [30]. According to the SIL:

“The initiative aims to promote analysis in artificial intelligence, science, technology, and innovation, as well as their understanding, knowledge, development, learning mechanisms, and responsible application and use.”

However, the text of the initiative focuses on amending subparagraph IV of Article 30 of the General Education Law, which would read as follows:

“Article 30. The content of the curricula and programs of study offered by the State, its decentralized entities, and private educational institutions, whether authorized or officially recognized, according to the type and level of education, shall include, among other things: (…)

IV. The promotion of research, analysis in artificial intelligence, science, technology, and innovation, as well as their understanding, knowledge, development, learning mechanisms, and responsible application and use (…).”

The fourth initiative (by the PRI political party) aims to make the teaching of computer skills and the use of artificial intelligence tools mandatory in the curricula of basic and upper secondary education [31]. To this end, it amends Articles 16, 18, 30, and 85 of the General Education Law. It incorporates a digital and technological orientation into the education system, emphasizing the development of digital competencies, including computational thinking, programming skills, and active participation in digital environments. They include studying artificial intelligence and its ethical and responsible use as part of technological knowledge, and using information, communication, and digital learning technologies to support logical thinking, problem-solving, innovation, and critical engagement across different educational levels.

The provisions also promote research, science, technology, and innovation by developing data analysis, programming, and artificial intelligence competencies at the educational level. In addition, they make educational authorities responsible for promoting teacher training and professional development in digital technologies, artificial intelligence, and programming, and for encouraging curricular and extracurricular activities related to digital culture and the ethical use of artificial intelligence.

A modification of the General Law on the Rights of Girls, Boys, and Adolescents, including the definition of artificial intelligence (in Article 4. XVI bis), is also proposed, through the modification of Article 58 (...).

“XII. To promote computational thinking and the development of programming skills and active participation in digital environments, as well as the study of artificial intelligence and its ethical and responsible use.”, and including artificial intelligence in articles 59 Bis and Ter.

2.3. Health-Related Initiatives

Four initiatives were identified in the search. Of these, one (presented twice) focuses on adding various provisions to the General Health Law, while another initiative concentrates only on Article 17 Bis of that Law. The third one focuses on Neurorights, and the last one on consumer protection.

The draft decree that adds various provisions to the General Health Law regarding artificial intelligence was presented by Ramírez-Barga of the PAN political party twice, first on 28 August 2024 [32] and then on 5 December 2024 [33], with the same text.

The initiative aims to regulate the use of artificial intelligence applied to health. To this end, it amends Articles 6, 7, 17 Bis, 53 Bis, 103 Bis 3, 103 Bis 8, 103 Bis 18, and 109 Bis of the General Health Law. The provisions describe the incorporation of artificial intelligence and information and communications technologies into healthcare services to expand coverage and improve quality. Artificial intelligence in health is defined as computational systems that use algorithms, machine learning, and data processing to perform tasks such as collecting and analyzing data, supporting diagnoses, recommending treatments, predicting outcomes, and contributing to the development of therapies and medications within the National Health System.

The text establishes conditions for the use of artificial intelligence in healthcare, including requirements related to reliability, precision, security, quality, therapeutic efficacy, and the protection of sensitive personal data. The use of these systems requires informed patient consent, anonymization, and the confidential handling of data used for training, and the possibility for patients to request explanations of AI-assisted medical decisions, their limitations, and the data on which they are based.

The proposal also outlines regulatory, supervision, and responsibility arrangements. Developers and providers must register AI systems with the health authority, and only systems authorized by the Ministry of Health may be used. The Ministry is responsible for regulation, evaluation, and auditing, and for determining when systems should be suspended or canceled due to health risks. Healthcare professionals, service providers, and AI developers are assigned legal and professional responsibility for outcomes associated with the use of artificial intelligence systems.

After the presentation of the initiative to the Health Commission of the Deputies Chamber, it was deemed to be “pending in commissions”, a status that it maintains to date.

The third health-related initiative (of the PV political party) is discussed below. The draft decree establishes the General Law on Neuro-Rights and Neurotechnologies, amending and adding various provisions related to neuro-rights and neurotechnologies [34]. This initiative focuses on human rights related to the central and peripheral nervous systems, brain and mental activity, and the information derived from them. The initiative addresses AI-related issues only in Chapter III, which deals with the integration of neurotechnologies with artificial intelligence and big data systems.

Chapter III comprises four articles. The first two focus on the use of AI under “strict ethical standards, guaranteeing transparency, integrity, non-discrimination, reliability, accountability, respect for privacy, identity, personal autonomy, and security”, and on obtaining informed consent from individuals before the use of neurotechnologies integrated with artificial intelligence and big data systems.

The remaining two articles focus on regulating entities that develop and operate artificial intelligence and big data systems that use neurotechnologies to provide the proposed Commission on Neuroethics and Neurorights with clear, accessible information about their operations, including the algorithms and the data sources used.

According to the initiative, this commission will be responsible for establishing mechanisms for ongoing monitoring and evaluation, and the results must be transparent and accessible to the public.

Finally, an initiative by Rendon-Gómez of the MORENA political party [35] aims to empower COFEPRIS (the Mexican Federal Commission for the Protection against Sanitary Risks). To this end, it proposes that the commission be granted the authority to incorporate innovative technologies, including artificial intelligence or technologies that improve upon or surpass it, in the performance of its technical functions, in order to reduce processing times and improve services, provided that their use is carried out under human supervision and in compliance with the principles of legality, proportionality, and transparency.

2.4. Intellectual Property Initiatives

An initiative of the PT political party was identified that amends and adds provisions to the Federal Cinematography Law, the Federal Copyright Law, and the Federal Labor Law, establishing protective measures for voice actors in Mexico against the use of artificial intelligence or emerging technologies in dubbing processes [36]. To this end, it modifies Articles 10 and 23 of the Federal Cinematography Law; 16, 21, 27, 113, and 113 Bis of the Federal Copyright Law; and 25, 132, 304, 304 Bis, and 994 of the Federal Labor Law.

The initiative aims to protect voice actors in Mexico from the misuse of artificial intelligence and emerging technologies in dubbing and audiovisual production. It requires producers to comply with labor, copyright, and performers’ rights laws to ensure ethical AI use, protect jobs, and obtain performers’ informed consent. It also prioritizes dubbing work in Mexico and favors the participation of living Mexican voice professionals to preserve national linguistic and cultural identity.

The proposal establishes that the cloning, synthetic reproduction, emulation, or modification of a performer’s voice using AI is prohibited without prior, informed, and written consent, as well as fair and proportional compensation. Both moral and economic rights holders may authorize or object to such uses, and any AI-generated or AI-assisted voice performance must clearly disclose the use of artificial intelligence to the public. Unauthorized AI voice use is explicitly excluded from being considered a valid dissemination of an artistic performance.

Additionally, the initiative strengthens labor protections by recognizing voice acting as an employment relationship, even for project-based or intermittent work. Employment contracts must include explicit clauses addressing the use of AI technologies, detailing consent, duration, purpose, and compensation. Unauthorized AI voice replication is considered a violation of labor and cultural rights. It is subject to significant fines, while the use of some AI-generated voices contributes financially to a fund managed by the Ministry of Culture to support the professionalization of the national creative industries.

2.5. Justice-Related Initiatives

Below, we address initiatives to regulate AI in several criminal laws. Five of those (including one with two different numbers) focused on changes to the General Law on Women’s Access to a Life Free of Violence.

Vences-Valencia of the MORENA political party proposed an initiative to amend and add articles 20 Ter and 20 Quater of the General Law on Women’s Access to a Life Free of Violence [37]. In the case of Article 20 Ter, Section IX is added.

Article 20 Ter. Political violence against women can be expressed, among others, through the following behaviors (…)

IX.

To defame, slander, insult, or make any expression that denigrates or disqualifies women in the exercise of their political functions, including the creation and dissemination of defamation campaigns or the generation of fake social media accounts or profiles, based on gender stereotypes, with the objective or result of undermining their public image or limiting or nullifying their rights.

X.

Disseminating images, messages or private information of a female candidate or in office, as well as manipulating and disseminating images or sounds created or modified through the use of artificial intelligence, or that use contextualized data or statements, by any physical or virtual means, with the purpose of discrediting, defaming, denigrating her and calling into question her capacity or skills for politics, based on gender stereotypes.

In the case of Article 20 Quater, it currently reads:

“Article 20 Quater. Digital violence is any intentional action carried out through the use of information and communication technologies, by which real or simulated images, audios or videos of intimate sexual content of a person are exposed, distributed, disseminated, exhibited, transmitted, marketed, offered, exchanged or shared without their consent, approval or authorization and that causes them psychological or emotional harm, in any area of their private life or in their self-image.

As well as those malicious acts that cause harm to the intimacy, privacy, and/or dignity of women, committed through information and communication technologies.

For the purposes of this Chapter, Information and Communication Technologies shall be understood to mean those resources, tools, and programs that are used to process, manage, and share information through various technological media.

Digital violence will be punished in the manner and terms established by the Federal Penal Code”.

The proposed modification [37] consists of expanding the second paragraph and modifying the third, as follows:

“Article 20 Quater. (…) Digital assassination will also be considered part of digital violence, which consists of all those actions carried out by oneself or through an intermediary, consisting of the creation of fake social media accounts or profiles, the creation and dissemination of defamation campaigns, the theft and alteration of personal data or the impersonation of a woman, with the purpose of slandering, intimidating or terrorizing women, and which are carried out through information technologies, digital platforms, social networks or the use of artificial intelligence.

For the purposes of this Chapter, Information and Communication Technologies shall be understood to mean those resources, tools, and programs, including those that use artificial intelligence, that are used to process, manage, and share information through various technological media.”

González-González of the PRI political party also tried to amend Article 20 Quater of the General Law on Women’s Access to a Life Free of Violence [38]. The initiative proposes to reform the third paragraph of Article 20 Quater of the General Law on Women’s Access to a Life Free of Violence, to read as follows:

Article 20 Quater. Paragraph 3. “For the purposes of this Chapter, Information and Communication Technologies shall be understood as those resources, tools, programs, and artificial intelligence platforms that are used to create, process, manage, share, and disseminate information through various technological means.”

It should be noted that in the SIL, this proposal appears twice, having two subject numbers: 4,810,351 and 4,810,368, and being found on two different URLs: https://sil.gobernacion.gob.mx/Librerias/pp_ReporteSeguimiento.php?SID=50b4db49742644687cedaccfd7c2d102&Seguimiento=4811102&Asunto=4810351 (accessed on 16 February 2026); https://sil.gobernacion.gob.mx/Librerias/pp_ReporteSeguimiento.php?SID=50b4db49742644687cedaccfd7c2d102&Seguimiento=4811106&Asunto=4810368 (accessed on 16 February 2026).

The same article is central to the initiative proposed by Monreal-Avila, also from the MORENA political party, to amend and add to Article 20 Quater of the General Law on Women’s Access to a Life Free of Violence [39]. His initiative adds paragraphs 5 and 6 to Article 20 Quater, as follows:

“In addition to what is established in the previous paragraph, digital violence will be considered to include the creation, distribution, dissemination, exhibition, transmission, commercialization, offering, exchange or sharing of images, audios or videos that have been falsified or manipulated through the use of artificial intelligence or any other technology, in order to damage the reputation, privacy or integrity of a person, without their consent or authorization.

For the purposes of this Law, artificial intelligence shall be understood as any computer system or program that uses algorithms or machine learning techniques to process and analyze data, and that is capable of generating digital content, including images, audio, or video, that are indistinguishable from real ones.”

Another proposal focused on amending Article 5 and adding Article 20 Septies to the General Law on Women’s Access to a Life Free of Violence, regarding generative artificial intelligence by Gómez-Cárdenas of the PAN political party [40].

“Article 5. For the purposes of this law, the following definitions shall apply: (…) XVIII. Generative artificial intelligence: A branch of artificial intelligence dedicated to creating entirely new data or content from existing data, using machine learning algorithms. Artificial intelligence is a branch of computer science that focuses on designing and building systems capable of performing tasks associated with human intelligence (…).”

In addition, Article 20 Septies is added:

“The use of generative artificial intelligence to create, modify, alter, generate, add and/or produce, in whole or in part, real or simulated images, audios or videos of intimate sexual content of a person without their consent, approval or authorization, and which generates any type of violence typified in this law, in any area of their private life or in their own image, is considered digital and media violence and will be sanctioned in accordance with the applicable criminal provisions.”

Another proposal by the PAN political party focused on amending various provisions of the Federal Penal Code, the General Law to Prevent, Punish, and Eradicate Crimes Related to Trafficking in Persons and to Protect and Assist Victims of these Crimes, and the General Law on Women’s Access to a Life Free of Violence [41]. This proposal expands the penalties for crimes related to pornography, human trafficking, and violation of sexual privacy, explicitly integrating artificial intelligence applications as a means of committing the crime and establishing sanctions that range up to 30 years in prison. It intends to modify the Federal Penal Code as follows:

“Article 199 Octies. The crime of violation of sexual privacy is committed by any person who willfully possesses, for purposes of commercialization or free or onerous distribution, discloses, shares, distributes or publishes images, videos or audios of intimate or sexual content of a person of legal age, without their consent, approval or authorization, real, simulated, elaborated or edited, or a combination thereof, through any information technology, including artificial intelligence tools or applications.

As well as anyone who videotapes, audiotapes, photographs, prints, creates, or edits using any information technology, including artificial intelligence tools or applications, images, audios, or videos with intimate sexual content of a person without their consent, approval, or authorization.”

The initiative also aims to amend the General Law to Prevent, Punish, and Eradicate Crimes Related to Trafficking in Persons and to Protect and Assist Victims of these Crimes, as follows:

“Article 14. Anyone who subjects a person or benefits from subjecting a person to perform pornographic acts shall be punished with a prison sentence of 10 to 15 years and a fine of one thousand to thirty thousand Units of Measurement and Updating (UMA). For the purposes of this article, the use of information and communication technologies, such as applications, tools, programs, platforms, and devices, including those using artificial intelligence, used to process, manage, edit, disseminate, or create content with information or material obtained from a victim, shall also be considered.

Article 15. Anyone who profits economically from the exploitation of a person through the trade, distribution, exhibition, circulation or offering of books, magazines, writings, recordings, films, photographs, printed advertisements, images or objects of a lascivious or sexual nature, real, elaborated, simulated or edited by any information technology, including tools or applications of artificial intelligence or a combination thereof, whether physically or through any means, shall be punished with a prison sentence of 5 to 15 years and a fine of one thousand to 30 thousand Units of Measurement and Update.

Article 16. A penalty shall be imposed (…) on anyone who procures, promotes, compels, advertises, manages, facilitates or induces a person (…) to perform sexual acts or acts of bodily exhibitionism, for sexual purposes, real or simulated, with the objective of producing videos, audios, photographs, films, real, elaborated, simulated or edited by means of any information technology, including tools or applications of artificial intelligence or a combination thereof (…).”

Reforms to the General Law on Women’s Access to Life Free of Violence are also proposed, as follows:

“Article 20 Quarter. Digital violence is any intentional act carried out through the use of information and communication technologies, by which real, elaborated, simulated or edited images, audios or videos are exposed, distributed, disseminated, exhibited, transmitted, marketed, offered, exchanged or shared using any information technology, including artificial intelligence tools or applications or a combination thereof, of intimate sexual content of a person without their consent, approval or authorization, and which causes psychological or emotional harm, in any area of their private life or to their self-image. As well as those malicious acts that cause harm to the intimacy, privacy, and/or dignity of women, committed through information and communication technologies, including artificial intelligence tools or applications.”

Another initiative by the MORENA political party, which amends various provisions of the Federal Penal Code and the General Law on Women’s Access to a Life Free of Violence, regarding the classification of digital and media violence [42] includes artificial intelligence in digital violence (Articles 7, 14, and 20). In addition, it considers modifications to Articles 38 and 41 to ensure digital platforms, information technologies, and artificial intelligence do not promote violence against women. It also include the following responsibility to the Attorney General’s Office of the Republic (Article 47 I):

“(e) In the area of digital platforms, information and communication technologies, and artificial intelligence, as sites susceptible to being used to perpetrate violence against women, and declares that the federative entities are responsible, in accordance with the provisions of this law and the applicable regulations in the matter (Article 49 XXII): (d) To know about digital platforms, information and communication technologies and artificial intelligence, as sites that can be used to perpetrate violence against women.”

It intends to add storage (in physical or digital devices) as a crime in Article 199 Octies of the Federal Penal Code, as well as the use of digital platforms or AI to record video, audio, photographs, prints or create images, audio or videos with intimate sexual content of a person without their consent, without their approval, or without their authorization.

The proposal also considers aggravating the following consequences, effects, or impacts of a crime of the violation of the victim’s privacy information:

Article 199 Decies:

(a)

Psychological and emotional suffering: depression, anxiety, stress, fear, panic attacks.

(b)

Physical: incitement to assault another victim, appearance of pain in different parts of the body, or suicide.

(c)

Social isolation: permanent or temporary withdrawal from public, family, and social life.

(d)

Economic consequences for victims and their families: job loss, payment of legal fees, online protection services, or long-term treatment for mental illness or sexual health problems.

(e)

Limited mobility in online and/or offline spaces: abandonment of a particular internet platform or social network, change in physical address due to threats of aggression on digital networks.

Following the same article (199) of the Federal Penal Code, the initiative of MORENA [43] aims to include the manipulation of images for sexual purposes as a crime, equating its penalties with those established for those who solicit or distribute explicit sexual content of minors or vulnerable people. To do so, it modifies section Septies as follows:

“Article 199 Septies. (...) The same penalties shall be imposed on anyone who, using images of other people, digitally manipulates them, through artificial intelligence or other computer tools, to alter them for the purpose of making them appear naked or involved in explicit sexual activities, and distributes, markets, or stores them with knowledge of the fact.”

Regarding Article 199 Octies, we identified two proposals for modification: [44] from the PT political party and [45] from the PV political party. Article 199 Octies currently states:

“The crime of violation of sexual privacy is committed by any person who discloses, shares, distributes, or publishes images, videos, or audios of intimate sexual content of a person of legal age, without their consent, approval, or authorization. Likewise, anyone who videotapes, audiotapes, photographs, prints, or creates images, audios, or videos with intimate or sexual content of a person without their consent, approval, or authorization.”

The first one aims to add the following text:

“Likewise, the use of techniques, applications, or artificial intelligence programs for the creation, manipulation, and distribution of videos, audios, images, and prints with intimate sexual content of a person without their consent, without their approval, or without their authorization.”

The second one aims to add:

“It will be considered a crime to modify, alter, or disseminate photographs, videos, or audios with artificial intelligence, as well as to sell or obtain any economic benefit from said content without the consent or authorization of the person.”

Another initiative of the PV political party [46] aims to equate deepfakes with sexual abuse. To this end, it proposes adding Article 199 Undecies as follows:

“Anyone who distributes, transmits, delivers, publicly displays, or uses other means to allow others to view synthetic sexual images of people outside their original context, generated by computer synthesis, artificial intelligence, or other technological methods, shall be considered to have committed sexual abuse and shall be punished with six to ten years’ imprisonment and a fine of up to two hundred days’ wages.”

From a different perspective, the use of artificial intelligence for identity theft is penalized [47]. This initiative by the PRI political party proposes adding Article 252 Bis to the Federal Penal Code to define the crime of identity theft and penalize those who commit it, including the use of artificial intelligence. To this end, it proposes: (1) establishing that anyone who, without the consent of the natural or legal person, assumes, appropriates, transfers, or uses the identity of another, with the purpose of causing harm, damage, or obtaining an undue profit for themselves or a third party, will be committing the crime of identity theft; (2) determining that a prison sentence of two to five years and a fine of 200 to 1000 UMA (Unit of Measurement and Update) will be imposed, in addition to reparation of damages; and (3) defining artificial intelligence software as any system, device, or computer program that has the capacity to perform activities that traditionally require human intelligence, including, among others, natural language processing systems, machine learning, neural networks, and advanced algorithms.

Another initiative from the MORENA political party focuses on sanctioning the creation of false audio for public personalities generated by artificial intelligence (AI). To this end, it amends and adds Chapter III and articles 211 Ter and 211 Ter 1 to the Federal Penal Code [48].

Chapter III is entitled “Creation and Distribution of Fake Audios of Famous People using Artificial Intelligence”, and Article 211 Ter. defines fake audio as:

“Any sound recording generated or manipulated by artificial intelligence with the intention of making it appear to come from a specific person, without their consent” and a public person as “an individual who enjoys significant public recognition due to his or her participation in public life, including, but not limited to, celebrities, politicians, athletes, and entertainment figures.”

Then, Article 211 Ter 1 considers that:

“The following will be sanctioned for the creation and distribution: (a) The creation, manipulation, distribution, and dissemination of false audio recordings of public personalities, using artificial intelligence or other technologies with the purpose of deceiving, defaming, or harming, is prohibited.”

Regarding the use of AI in law enforcement, the PT political party initiative [49] adds Article 8 Bis to the Federal Law against Organized Crime. The proposal authorizes public security institutions to use artificial intelligence to investigate, identify, and analyze criminal activities by organized crime.

This includes pattern analysis, real-time monitoring and analysis of communications and financial networks, implementation of biometric recognition systems, risk assessment in financial operations, automation and optimization of criminal investigation processes, identification of criminal modus operandi, and creation of AI-optimized inter-institutional databases, among other activities.

The implementation and operation of artificial intelligence must be supervised by the Ministry of Security and Citizen Protection to ensure its proper use and prevent potential abuses, in accordance with applicable legal provisions. Security protocols will be established to protect the processing and handling of information and products generated by artificial intelligence, preventing misuse of the technology and potential privacy breaches.

Finally, the PAN political party has launched an initiative to amend and add various provisions to the Federal Penal Code and the Federal Copyright Law, regarding generative artificial intelligence [50]. The initiative aims to combat the misuse of generative artificial intelligence. To this end, it amends Articles 211 Bis 8-10 of the Federal Penal Code and Articles 4, 14, and 101 of the Federal Copyright Law.

It considers what will be understood as misuse of technological tools based on generative artificial intelligence:

(a)

market introduction, commissioning or the exploitation thereof that results in a deliberate, deceptive and non-consensual manipulation to alter or modify images, scenarios or video and audio recordings, making them appear real and to cause prejudice and damage to a person or a group of people and/or diminish the capacity of a person or a group of people to make an informed decision, seeking to make them adopt a position or criterion or make a decision that they would not otherwise have made;

(b)

the introduction into the market, the putting into service or the exploitation of technological tools of generative artificial intelligence that are fed by facial recognition data through the non-selective extraction of facial images from the internet or closed circuits as well as data that may be considered private or personal, or that make illicit use of remote biometric identification and information systems or that falsify official information or documents;

(c)

the introduction into the market, the putting into service, or the exploitation of technological tools of generative artificial intelligence through which confidential or reserved information, or images, videos, or audios, illegally extracted without the consent of the affected parties, and are disseminated or spread;

(d)

the use and exploitation of technological tools of generative artificial intelligence to generate content and works that, without specifying that they have been developed through the use of them, are presented as their own and seek to be registered as such for their commercialization or that infringe existing rights of protected works or that result in the commission of a fraud to the detriment of any natural or legal person.

It also stipulates that anyone who commits any of the acts related to the misuse of technological tools based on generative artificial intelligence will be sentenced to six months to four years in prison and a fine of one hundred to six hundred days.

Regarding copyright, it determines that content and systems created by technological tools based on generative artificial intelligence that infringe existing rights of prior works or protected original creations, that do not have authorization for the use of the elements incorporated in the work, that do not provide the inputs for their creation, that are generated autonomously and that do not make transparent their sources, the technology used, their respective processes and stages of creation and the authors or co-authors involved in it, will not be subject to copyright protection. It also considers AI-generated computer programs to be computer programs (which are suitable for copyright under current legislation).

2.6. Promotion of AI Initiatives

Three legislative initiatives were found, two of which were proposed by Congressman Rodríguez-Carrillo of the MC political party [51,52] to amend the Science and Technology Law.

The first one (2021) changed Article 2 to include ethics and human rights and the promotion of AI in the foundations of a state policy to support the integration of the National System of Science, Technology, and Innovation, as follows [51]:

Article 2. (...)

IX.

To promote the development of an ethical and human rights framework that governs national science, technology, and innovation policy;

X.

To promote the use of artificial intelligence to solve fundamental national problems, contribute to the country’s development, and improve the well-being of the population in all aspects, with strict adherence to and respect for human rights.

Then it defines artificial intelligence in Article 4, as:

XV.

Artificial intelligence: any system that exhibits intelligent behavior, by being able to analyze its environment and take action with a certain degree of autonomy, to achieve specific objectives.

It also proposes including the President of the National Human Rights Commission on the General Council for Scientific Research, Technological Development, and Innovation. It considers that the General Council “shall have the following powers: (...) XII. To establish national guidelines on ethics related to science, technology, and innovation.”

Although this proposal is still pending, another version [52] was presented in 2024 (possibly due to the approval of a new General Law on Humanities, Sciences, Technologies, and Innovation in 2023). The new version “aims to promote the use of artificial intelligence to contribute to the country’s development and improve well-being, with strict adherence to human rights, and proposes to amend Articles 4, 11, 68, and 69 of the General Law on Humanities, Sciences, Technologies and Innovation”.

However, artificial intelligence is only mentioned twice: first, in the definition of artificial intelligence, added to Article 4, item XV Bis:

“Artificial intelligence: any system that exhibits intelligent behavior, by being able to analyze its environment and take action with a certain degree of autonomy, to achieve specific objectives.”

And subsequently, it is included in the addition of item XXVIII to Article 11, which states:

“To promote the use of artificial intelligence to solve fundamental national problems, contribute to the country’s development, and enhance the well-being of the population in all aspects, with strict adherence to and respect for human rights.”

Similarly to priors, the third initiative [53] (of MORENA political party) focuses on amending and adding articles 4 and 63 to the General Law on Humanities, Sciences, Technology, and Innovation. The proposal defines artificial intelligence as “systems designed to operate with varying levels of autonomy, capable of adapting, and that perform tasks based on explicit or implicit instructions, or by inferring from the input it receives. These tasks can involve making predictions, providing recommendations, or making decisions, which can be compared to the reasoning processes of the human mind.” (Article 4).

It also modifies Article 63 by adding fractions XXVIII and XXIX as follows:

XXVIII.

Promote the development, use, and application of artificial intelligence (AI) technologies as part of the national policy on science, humanities, technology, and innovation, in accordance with the principles of ethics, inclusion, security, data protection, sustainability, and human rights.

XXIX.

Develop, integrate, update, and approve a National Ethical Framework for the use and development of artificial intelligence, in accordance with the principles of this law.

2.7. Regulate AI Initiatives

We identify four general-purpose initiatives to regulate AI in Mexico. All were presented in late 2024 and early 2025. Below, we summarize the initiatives in chronological order.

The first proposal (3 October 2024, by PAN political party) adds a section XXIII Bis to article 4 and an article 101 Bis 3 to the General Law on the Rights of Girls, Boys, and Adolescents [54], to regulate the use of AI tools for biometric identification of minors. Article 4. XXIII bis defines facial recognition, and Article 101 Bis 3 proposes that the use of AI be permitted when searching for missing persons.

The article states that informed consent from legal guardians is required for the use of biometric data, except when the law permits its use without such consent in cases of urgency or security. The use of these technologies will be monitored and audited periodically by a specialized committee composed of experts in children’s rights, technology, and data protection, to ensure their correct application and the comprehensive protection of the rights of girls, boys, and adolescents.

Furthermore, the article enables the competent authorities to promote inter-institutional collaboration with other government entities, international organizations, the private sector, and non-governmental organizations, as needed, to share relevant information, technological resources, and best practices for the use of artificial intelligence technologies.

The second proposal, presented on 12 November 2024, by the MORENA political party, focused on amending and adding various articles of the current Federal Telecommunications and Broadcasting Law [55]. The initiative aims to ensure that artificial intelligence used in telecommunications services is used responsibly and ethically:

“Article 7. With regard to Artificial Intelligence Systems, the Institute will establish harmonized standards (…). These guidelines will apply to suppliers that introduce or put into service Artificial Intelligence systems in Mexico, and to suppliers and users of Artificial Intelligence systems located in another country when the output information generated by the system is used within the national territory.”

It emphasizes that the implementation and use of artificial intelligence must be based on the principles of full respect for human rights, the preservation of life and human dignity, transparency, the protection of personal data, and strict adherence to social ethics. Developers and implementers of artificial intelligence will at all times maintain a decryptable cryptographic file to ensure the content created by the artificial intelligence remains distinguishable.

In addition, artificial intelligence systems must be designed and implemented with full respect for the intellectual property and copyrights of the creators of the data and algorithms used in their development, and ensure transparency at all times regarding their operation and the results generated by the system’s use. These artificial intelligence systems must not discriminate against people based on race, gender, religion, sexual orientation, disability, or any other reason, creating a safe environment that fully respects the privacy of individuals.

It holds that developers and operators of artificial intelligence systems will be responsible for any damage arising from their operation. To generate a clear distinction between content generated by artificial intelligence systems, these systems must have a cryptographic digital fingerprint that allows identification of their origin and operation.

The article 190 Ter summarizes prohibited artificial intelligence practices, as follows:

(a)

The introduction into the market, the putting into service, or the use of an Artificial Intelligence system that uses subliminal techniques that transcend a person’s consciousness to substantially alter their behavior in a way that causes or is likely to cause physical or psychological harm to that person or to a third party.

(b)

The introduction into the market, the putting into service or the use of an Artificial Intelligence system that takes advantage of any of the vulnerabilities of a specific group of people due to their age or physical or mental disability in order to substantially alter the behavior of a person belonging to that group in a way that causes or is likely to cause physical or psychological harm to that person or to another person.

(c)

The introduction into the market, the commissioning or the use of Artificial Intelligence systems by the authorities or on their behalf for the purpose of assessing or classifying the reliability of natural persons during a specified period of time based on their social conduct or known or predicted personal or personality characteristics, such that the resulting social classification causes one or more of the following situations:

(1)

Prejudicial or unfavorable treatment towards certain individuals or entire groups in social contexts that are unrelated to the contexts where the data were originally generated or collected;

(2)

Harmful or unfavorable treatment towards certain individuals or entire groups that is unjustified or disproportionate with respect to their social behavior or the seriousness thereof.

(d)

The use of remote (real-time) biometric identification systems in publicly accessible spaces for law enforcement purposes, except to the extent that such use is strictly necessary, justified, and motivated before the relevant Authority to achieve one or more of the following objectives:

(1)

The selective search for specific potential victims of a crime, including missing minors;

(2)

The prevention of a specific, significant, and imminent threat to the life or physical safety of individuals or a terrorist attack. Any specific use of a real-time remote biometric identification system in a publicly accessible space for law enforcement purposes will be subject to prior authorization from a judicial or relevant administrative authority, which will grant such authorization upon a reasoned and substantiated request for its implementation. However, in a duly justified emergency situation, the system may be put into use before obtaining the corresponding authorization, which may be requested during or after its use. The competent judicial or administrative authority will only grant authorization when it is convinced, based on objective evidence or clear indications presented to it, that the system is justified.

The third initiative was presented the following month (13 December 2024 by PV political party), proposing a National Law that regulates the use of artificial intelligence [56]. According to Article 3, the purposes of this Law are:

• The establishment of the national policy for the use and application of Artificial Intelligence through the various mechanisms that facilitate immediate, prompt, and real-time access for people.
• The use and application of Artificial Intelligence in information and communication technologies as a linking tool.
• Establishing the procedures, instances, and mechanisms in which Artificial Intelligence will be applied.
• Define and establish the regulatory bodies, as well as the strategies for the use and exploitation of Artificial Intelligence.
• Promote the use of technologies under which Artificial Intelligence is applied, except for those established by applicable legislation.

AI is defined as “The scientific discipline that deals with the creation of computer programs that perform operations comparable to those performed by the human mind, similar to learning and logical reasoning,” which is considered a good definition.

The creation of three bodies that will intervene in the country’s AI policies is proposed:

• The National Artificial Intelligence System in Article 18 (articulating the mechanism of AI actions that will establish and execute the National Artificial Intelligence Strategy and the actions, policies, and programs of a specific and sectoral nature. It is made up of the Council; the authorities; and the social, private, and academic sectors.
• The National Council for Artificial Intelligence in Article 21 (the highest body for coordination and definition of public policies in this area, includes the presidency of the federal executive branch and the heads of the Secretariats of the Interior, National Defense, Navy, Public Function, Security and Citizen Protection and Education, as well as the heads of the Chief Administrative Office and the National Commission for Regulatory Improvement)
• The National Center for Artificial Intelligence in Article 13 (decentralized body of an operational nature and coordinator of actions for the implementation of the national artificial intelligence policy)

It establishes 18 technological standards (Article 31): accessibility, openness, technological suitability, confidentiality, preservation, availability, efficiency, balance, equality, integrity, accountability, maximum transparency, privacy, usability, proper handling of information, ethical use of information, and information quality. It also states that:

“Systems, interfaces, programs, and applications that use artificial intelligence must be transparent, traceable, environmentally friendly, and subject to continuous human oversight; therefore, they must not be fully automated under any circumstances” (Article 42).

It emphasizes the privacy of individuals, considering that “private communications are inviolable” and that “only the judicial authority, at the request of the authority empowered by law or the head of the Public Ministry, may authorize the intervention of any private communication“. (Article 35).

The initiative states that applications involving processes, procedures, and services that require AI intervention or contemplate AI model integration must be carried out under the following parameters (Article 40).

It is established that the public sector must “inform the user in advance of its use and have their express consent” (Article 44) and may not use AI for (Article 45):

• Processes that require collegial deliberation.
• Resolutions of an administrative contentious nature.
• Hearings and mechanisms for citizen participation and consultation.
• Medical diagnoses.
• The implementation of Alternative Dispute Resolution methods.
• Those whose resolution requires the presence of the authorities.

It also defines three types of risks in AI systems:

Article 47. Unacceptable risk is understood to mean Artificial Intelligence systems that pose a real, possible, and imminent threat to people and involve schemes for cognitive manipulation of behavior, classification of people based on their behavior and socioeconomic level, or personal characteristics.

Article 48. High risk is understood to mean Artificial Intelligence systems that have a significant potential to negatively affect citizens’ security, the protection of human rights, legality, and legal certainty, and whose application is the exclusive responsibility of Public Security corporations and the Armed Forces.

Article 49. Limited risk is understood to mean Artificial Intelligence systems for user interaction that allow users to make informed decisions; users’ consent regarding their interaction with Artificial Intelligence is expressed, and they may decide at any time and under any circumstances how to use it.

Regarding non-compliance with the law (Article 50), only infractions by “public servants of the various levels of government” are considered, and these will be sanctioned in accordance with “the General Law of Administrative Responsibilities.”

The fourth proposal (30 April 2025, by the MORENA political party) also aims to present a Federal Law for the Ethical, Sovereign, and Inclusive Development of Artificial Intelligence [57]. The text comprises 29 chapters that address various components of the AI ecosystem. These include general provisions and key definitions; the creation of a National AI Council; a proposed risk assessment system, a progressive regime of rights, duties, and sanctions, and mechanisms for oversight, redress, and algorithmic justice.

It also addresses strategic issues such as protecting children and adolescents from harmful content; strengthening regulation across sectors like health, land-use planning, justice, and education; and the obligation to ensure explainability, technical traceability, and the right of appeal against automated decisions. It also includes a specific chapter on digital sustainability and environmental impact.

Among the proposals, the creation of the National Council of Artificial Intelligence stands out, “as a decentralized public body of the APF, not sectorized, with its own legal personality and assets, technical and management autonomy, with its domicile in Mexico City” and the National Algorithmic Audit Platform (PNAA) as an autonomous technical body attached to the National Council of AI, “which will be the competent authority to evaluate, diagnose and issue technical opinions on data sets, models and artificial intelligence systems”.

The definition of Prohibited Artificial Intelligence is considered interesting, as it literally states:

“Prohibited artificial intelligence shall be understood to mean those systems whose development, training, implementation, use, commercialization, import, or export is prohibited because it represents an unacceptable risk to human dignity, fundamental rights, life, public safety, social peace, or democratic stability”.

Among others, artificial intelligence systems shall be considered to include those that:

(a)

Use subliminal manipulation techniques or extreme psychological persuasion to alter human behavior without informed consent. (…)

(e)

They are developed for the creation, improvement, or use of weapons for the exclusive use of armed forces, or that facilitate the development of chemical, biological, radiological, or nuclear weapons, prohibited by international treaties.

(f)

They are intended for the manufacture or dissemination of explosives, toxic agents, or substances intended to cause massive damage to persons or property. (…)

The ban will be applied

“broadly and preventively, in accordance with the principles of precaution, enhanced protection of human rights, and safeguarding of public safety.”

Article 63 regulates which artificial intelligence systems will be prohibited throughout the national territory:

• Generate, recommend, or disseminate content that promotes violence, self-harm, hatred, discrimination, or harmful behaviors;
• They stimulate digital addictions, technological dependence, or serious emotional disturbances;
  (…)
• Train artificial intelligence models using data from minors without the express, informed, and verifiable consent of the person exercising parental authority, custody, or guardianship.
  (…)

Furthermore, Article 66 establishes that:

“Every artificial intelligence system accessible to or directed at minors must have a Child Algorithmic Impact Assessment.”

The initiative includes the possibility of fines and even criminal prosecution:

“When the use of an artificial intelligence system: I. Causes injury, death, deprivation of liberty, serious property damage, or II. If part of criminal conduct such as fraud, electoral manipulation, digital violence, human trafficking, or illegal surveillance, the acts may be prosecuted criminally in accordance with the Federal Penal Code, without prejudice to administrative or civil sanctions” (Article 133).

In summary, we examined 38 initiatives undertaken by six political parties in Mexico (

Table 2. Summary of reviewed initiatives per area and political party.

Area	MC	MORENA	PAN	PRI	PT	PV
Congress	0	4	3	1	1	0
Education	0	2	1	1	0	0
Health	0	1	1	0	0	1
Intellectual Property	0	0	0	0	1	0
Justice	0	5	3	2	2	2
Promote AI	2	1	0	0	0	0
Regulate IA	0	2	1	0	0	1

). The findings suggest that the MC party’s interests are focused solely on promoting AI. PT is the only party concerned about the intellectual property of the Mexican actors and the foreseeable difficulties arising from the rise of generative AI for automatic voice generation. PV is interested in justice-related topics (particularly with regard to the use of deepfakes of a sexual nature), neurorights, and general AI regulation. PRI is interested in empowering Congress to regulate AI, introduce AI skills in education, and address issues of justice (fake social media profiles and identity theft). The MORENA and PAN parties are the most involved in AI-related laws, with broader interests (considering all areas apart from intellectual property).

3. Discussion

In Section 2, we answered our first research question: Q1: What legislative initiatives currently exist in Mexico that directly or indirectly address the use of artificial intelligence? Therefore, this section focuses on our second question: Q2: What critical aspects of AI (such as algorithmic bias, transparency, accountability, or privacy) are insufficiently regulated or completely absent in the current legislation? First, we state the limitations of our study, and in Section 3.2, we summarize the main gaps found per area.

3.1. Limitations of the Study

The main limitations of our study are:

6.

Documentary and normative scope: The analysis is limited to formal legislative initiatives, excluding regulations, administrative guidelines, case law, and informal regulatory practices that also influence AI governance in Mexico.

7.

Static nature of the analysis: Since the study is based on pending initiatives at a specific point in time (31 May 2025), the results may be affected by the rapid evolution of the legislative and technological landscape, which limits the timeliness of the conclusions.

8.

Textual search: The Legislative Information System search engine only allows for textual search; that is, only the words “artificial intelligence” (inteligencia artificial in Spanish) were used in the search.

9.

Lack of empirical validation: This study does not incorporate the perspective of key stakeholders (legislators, regulators, developers, or AI users), which restricts the empirical assessment of the practical impact and potential effectiveness of the initiatives analyzed.

10.

We only focused on the current legislative landscape regarding AI in Mexico. We did not compare it with those in other countries.

3.2. Main Gaps of the Surveyed Initiatives

Below, we address the main gaps identified in the proposed initiatives by area of interest.

3.2.1. Congress-Related Initiatives

With respect to initiatives to empower Congress to enact AI-related laws, only one addresses the prevention of AI-based discrimination at the constitutional level. In addition, there is no consensus among legislators regarding the modification of an (apparently simple) text that includes AI among the areas over which Congress has the power to enact laws.

3.2.2. Education-Related Initiatives

In the Education-related initiatives, there is no regulatory framework to ensure the necessary technological infrastructure [28], including data centers, high-speed networks, and hardware that meet standards to guarantee coverage and keep pace with AI advancements, as well as addressing data usage and algorithmic bias.

There is no clear explanation of how training programs [29] will be implemented, how discrimination and algorithmic bias will be prevented, or how job losses resulting from the implementation of AI to increase business productivity will be mitigated.

Finally, the wording of the proposal [30] is unclear, particularly regarding the aspect of “analysis in Artificial Intelligence.” This is because in Spanish grammar, the phrase “analysis in” typically refers to a specific field or context of application, which makes it difficult to determine how analysis within artificial intelligence will be incorporated into curricula across disciplines.

3.2.3. Health-Related Initiatives

With respect to the Health-related initiatives established by [32,33], it is not clear how “promoting” the protection of personal data will guarantee the security of such data, nor how AI will be incorporated into health services, whether it is to manage care, support diagnosis or treatment, record consultations (environmental scribes), or for what other purposes these technologies will be used. In one case, a set of potential applications is listed, but it is clarified that they are “not limited” to these.

It leaves data protection to individuals (the developers and providers of AI systems). No sensitive characteristics (such as age, sex, race, socioeconomic status, among others) are excluded from being used by developed AI algorithms. It leaves it to the patient to obtain “informed” consent without establishing who will provide accurate information about how the systems work (especially given the industrial secrecy usually associated with systems development).

The Federal Commission for the Protection against Health Risks assigned responsibility for assessing health risks associated with the use of artificial intelligence systems, without clarifying which risks are being assessed or how they will be evaluated. However, it will be the Ministry of Health that “will determine in which cases artificial intelligence systems pose a risk to people’s health and will proceed to request the suspension of the system’s operations and, where appropriate, its cancellation” without clarifying what mechanisms will be used for this purpose. This aspect is grave, given the negative impact that Large Language Models (LLMs) have on people with mental illness, increased phobias, paranoia, schizoid disorders, and a tendency towards violence and suicide, as well as the issues of hallucinations inherent to LLMs [58,59].

It is unclear how “equity and inclusiveness in its use” will be guaranteed, what type of training doctors will receive, or whether any requirements for AI systems are mandated. It dilutes the responsibility for events that may affect patients among “health professionals, health service providers, and developers of artificial intelligence systems intended for use in the health field, who will be responsible, as appropriate.”

Regarding the initiative [34], there is no regulation to ensure bias is absent, no assessment of system quality, and nothing that allows medical personnel or patients to determine the potential impact of these tools on their care, diagnosis, or treatment.

In the case of COFEPRIS, it is unclear which criteria will be used to determine an “improvement of services,” and no characteristics are excluded from the potential use of AI systems, which can introduce bias and disparities and increase inequity and algorithmic injustice.

3.2.4. Intellectual Property Initiatives

The voice-protection initiative [36] is a starting point for regulating the use of personal data (in this case, a person’s voice), but it fails to address crucial aspects such as algorithmic bias, transparency, technical accountability, and data protection. Furthermore, it presents implementation challenges in a global market and lacks alignment with international standards that would enable a flexible, comprehensive framework.

It focuses on protecting the rights of voice actors against AI-generated voice cloning, emphasizing labor and copyright issues, but neglecting technical aspects. It fails to define quality, transparency, or auditing standards applicable to the use of voice synthesis and cloning software. Similarly, it omits any form of regulation regarding the security of biometric data (such as voiceprints) used to train these models, thus exposing risks of data breaches, misuse, or unauthorized appropriation in international contexts.

In addition, it does not include mechanisms for auditing, monitoring, or mitigating potential biases in AI systems that could affect employment opportunities; for example, by favoring certain voices or profiles over others. Moreover, the initiative places the burden on employers and producers but does not clearly define the responsibility of AI developers, software distributors, or streaming platforms that use these technologies. While it establishes financial penalties for unauthorized uses in employment and commercial contexts, it is unclear how to address illicit non-commercial uses (e.g., parodies, viral content) that also affect artists’ identities and reputations. Finally, the requirement for authorization from the Ministry of Culture for certain AI uses could become a bureaucratic obstacle that stifles artistic experimentation or innovation in independent projects.

3.2.5. Justice-Related Initiatives

The main gaps found in the Justice-related initiatives are summarized below:

When considering the use of AI for generating or modifying sexually explicit content, some proposals fail to consider whether such content depicts a real person or if it is generated without reference to a particular individual.

Furthermore, some initiatives only consider malicious acts that cause harm to the intimacy, privacy, and/or dignity of women, or the creation and dissemination of defamation campaigns, the theft and alteration of personal data, or the impersonation of a woman, with the purpose of slandering, intimidating, or terrorizing women, thus rendering male victims of digital violence invisible and excluding them from access to justice.

The initiative that proposes equating any deepfake to sexual assault [46] effectively criminalizes the use of AI technologies for generating visual content that includes any sexual component, such as series and films, among others. The text does not consider whether a real victim exists. Still, it is ambiguous in its wording, referring to “synthetic sexual images of people outside their original context, generated through computer synthesis, artificial intelligence, or other technological methods.” This wording does not clarify what the “original context” of “synthetic images” is; such images are assumed to have been “generated through computer synthesis.”

The initiative that penalizes fake audio recordings [48] with the purpose of deceiving, defaming, or harming defines a fake audio recording as “Any sound recording generated or manipulated by artificial intelligence with the intention of making it appear to come from a specific person, without their consent.” In this way, the door is left open for audio recordings to be manipulated using any technology to deceive, with the individual’s consent, without violating the law. In addition, only fake audio recordings of “public figures” are penalized, and not those of any person.

The initiative to use AI to combat crime [49], despite stating that “security protocols will be established to protect the processing and handling of information and products generated by artificial intelligence, preventing the misuse of the technology and possible violations of privacy,” does not clarify who will guarantee due process, non-discrimination, and the presumption of innocence, among other relevant elements, since the proposal is “illustrative but not exhaustive” regarding the use of AI in biometrics, finance, behavioral prediction, and real-time monitoring of communications, among others. At no point is it stated that a court order, an open case, or anything similar is required to be subject to these AI applications, which, de facto, would enable real-time mass surveillance without legal checks and balances.

The introduction to the market and deployment [50] of generative AI platforms that result in deliberate, deceptive, and non-consensual manipulation to alter images, scenarios, or video and audio recordings, making them appear real, is penalized. This shifts responsibility for potential misuse of these tools to providers, not to users. Furthermore, many platforms that support image, audio, and video manipulation operate online. There is no guarantee of how their providers will be prosecuted.

The introduction to market, deployment, or operation of generative artificial intelligence technological tools that use facial recognition data obtained through the indiscriminate extraction of facial images from the internet is also penalized. This effectively penalizes any current generative AI platform (e.g., Gemini, Claude, ChatGPT, and others), since all of them have been trained with images from various sources, including the internet.

3.2.6. Promoting AI-Related Initiatives

Regarding the promotion of AI [51,52,53], it remains unclear how the use of AI is to be “promoted,” how “strict adherence to human rights” will be ensured, or under what conditions the use of AI will be considered acceptable or unacceptable. The initiatives also fail to address key technical aspects, including machine learning, algorithmic bias, transparency, privacy, and experimental validation, thereby limiting their potential to inform effective public policy.

3.2.7. Regulate AI-Related Initiatives

Regarding the initiative to modify the Federal Telecommunications and Broadcasting Law [55], we identified the following limitations. First, the international reach is difficult to control. This applies to providers and users outside Mexico whose systems affect people in Mexico, complicating oversight and enforcement. Second, it presents complex technical and transparency requirements: The obligation to generate cryptographic digital signatures and decryptable files of AI-generated content can be costly and technically challenging for many developers to implement. Developers and operators are the only ones liable for AI-related damages. Users are not considered liable for the misuse of the technology; only the developers and operators are. Third, there is ambiguity in ethical and human rights concepts: requirements such as “full respect for human rights” or “ethical principles” are vague and could lead to uncertainty regarding their practical interpretation.

The main limitations of the AI law of 2024 [56] are that it leaves the proposal of technical and methodological standards to the National Center, so in practice, it is unclear how compliance will be guaranteed, which could render regulations ineffective. It mentions principles (transparency, traceability, ethics, human oversight), but clear technical and methodological standards for ensuring their compliance are not defined, which may render them impractical.

Although risk categories are established (unacceptable, high, limited), verifiable technical criteria for classifying AI systems within each category or evaluation procedures are not included. In addition, there is no requirement to detect, mitigate, or report biases in models, which is a critical aspect in the scientific literature and international AI regulations. Moreover, the requirement to inform users and label AI-generated content is stated, but model explainability and the right to know how an automated decision was reached are not addressed. The source of the information to be cited [56] is not specified (Training information? Validation information? User information? Other?), and it is not clarified what constitutes “illegal content” that developers should avoid generating with the models. The obligation to prevent illegal content falls on the “developer,” without distinguishing between designers, implementers, distributors, and end users, creating gaps in accountability. Although consent is mentioned in the use of AI systems in the public sector, no details are provided regarding security measures, data retention, transfer, or anonymization for data used by AI systems.

Finally, there is no mention of what will happen as a result of non-compliance by individuals or legal entities that are not public servants of the various levels of government.

The AI law of 2025 [57] contains duplicate articles in the text (e.g., 67 and 69, 66 and 68). It does not clarify how or how often the “own funds” of the National Artificial Intelligence Council will be allocated (beyond the fact that the Federal Government will allocate them), nor how it will be ensured that the National Algorithmic Audit Platform can “maintain its technological capabilities, methodologies, and tools up to date, to preserve its effectiveness, reliability, and relevance to the state of the art in artificial intelligence and algorithmic auditing” (Article 21). Although Article 66 establishes that “Every artificial intelligence system accessible to or directed at minors must have a Child Algorithmic Impact Assessment”, it does not clarify how this law will be enforced, given that Mexican citizens (including minors) can use AI systems via HTTP protocol as long as they have the necessary internet connectivity. How do governments intend to regulate the use of these systems?

In addition, it does not specify how responsibility (Article 133) will be established, nor who will bear it—the developers, the operators, those who facilitate access to the systems, or other involved parties. Likewise, it does not clarify who will be held accountable under the law in case of non-compliance (the developers, companies, government entities, individuals who deploy the system, or others).

Although it includes a “broad and preventive” prohibition of some artificial intelligence systems, the writing of the text de facto prohibits, for example, professional training systems (e.g., military simulators developed for weapons use) by public and private entities responsible for security matters, as well as systems for detecting substances (because they could be harmful) used in scientific institutions for knowledge development. Furthermore, “subliminal manipulation” is not defined, leaving the door open to banning systems that include recommendations, advertising, and other elements that have a “filter bubble effect” [60,61,62] and could be considered behavior-altering. This would effectively classify all social media platforms as prohibited artificial intelligence systems.

Similarly, following the provisions of items II and VII of Article 63 of the proposal, all social networks would be effectively banned in this country, since they encourage digital addiction [63] and actively use the data of their users (many of whom are minors) for their operation (e.g., advertising) and the training of their models. For example, the minimum age for using Facebook in Mexico is 13.

Finally, no proposed initiatives align with emerging regulatory frameworks such as the EU AI Act or the OECD and UNESCO guidelines, which could lead to incompatibilities in cooperation or digital trade contexts.

4. Conclusions

In this paper, we aimed to answer two research questions: (Q1): What legislative initiatives currently exist in Mexico that directly or indirectly address the use of artificial intelligence? (Q2): What critical aspects of AI (such as algorithmic bias, transparency, accountability, or privacy) are insufficiently regulated or completely absent in current legislation?

We reviewed the current initiatives in the Mexican legislative bodies aimed at regulating artificial intelligence. We identified 40 pending initiatives across seven areas: Congress, Education, Health, Intellectual Property, Justice, Promote AI, and Regulate AI. Regarding the second, we can conclude that, in Mexico, the use of artificial intelligence (AI) faces several legal limitations, summarized as follows:

• All critical aspects are insufficiently regulated in the reviewed initiatives. There are specific provisions for bias-free training, no enforcement of transparency, insufficient accountability, and privacy is only covered by current data protection laws.
• Privacy and Data Protection: Existing privacy and data protection laws, such as the Federal Law on the Protection of Personal Data Held by Private Parties (in Spanish Ley Federal de Protección de Datos Personales en Posesión de los Particulares), apply indirectly to AI. However, they do not address all of this technology’s specificities.
• Ethics and Transparency: Implementing AI poses significant ethical challenges, including ensuring fairness, equity, and transparency in its use. There are currently no laws that address these aspects from a technical perspective.
• Oversight and Accountability: The algorithms and source code used in AI applications are not clearly supervised or regulated, which can lead to accountability and transparency issues. No current law requires code supervision or accountability for AI systems.
• Cybersecurity: The lack of a robust legislative framework on cybersecurity also affects the use of AI, as many AI applications depend on the security of the data and systems they operate.

Currently, Mexico lacks any approved legislation governing the use of AI. This allows providers and developers to use a wide range of data from multiple sources for model training and decision-making, leaving users potentially vulnerable to the unethical use of their personal data. This creates uncertainty and risks for both legal professionals and citizens.

These points reflect the need to develop a comprehensive regulatory framework that addresses AI’s specific challenges in Mexico. The regulatory landscape governing artificial intelligence (AI) in Mexico presents a complex interplay of existing legal frameworks and evolving proposals that aim to address the multifaceted challenges posed by AI technologies.

The need for comprehensive regulations, particularly given advances in cross-border technologies such as drone surveillance, underscores the urgency of a legal framework that recognizes national priorities and aligns with international obligations. This alignment is crucial, as it fosters a harmonized regulatory environment that can effectively manage the ethical and legal implications of AI integration while maintaining a human-centered perspective.

5. Challenges and Future Directions for AI Regulation in Mexico

An element of interest in Mexico is that there is currently no specialized framework that regulates the use of AI and protects citizens from discrimination, algorithmic bias [64], false representation, digital identity theft, or other practices that undermine their human rights, nor does it guarantee ethical AI governance. Therefore, a comprehensive legal strategy that addresses current challenges and positions Mexico to harness AI’s potential to drive innovation and social progress remains necessary.

Implementing AI regulations in Mexico presents numerous challenges, primarily due to the intricate interplay between legal, regulatory, and technical domains. The development of governance regimes for AI must intricately weave human rights, democracy, and the rule of law into a cohesive framework, requiring a comprehensive approach that recognizes these interconnections [65]. The technical challenges are further compounded by the need to establish robust regulatory frameworks that address privacy and security concerns, which are paramount given AI’s potential to disrupt existing norms across sectors such as medicine, finance, and law [66].

Furthermore, the introduction of new legislation must be balanced with efforts to reduce regulatory barriers to foster innovation while ensuring ethical compliance, which remains a significant challenge for countries such as Mexico, where emerging economies are struggling to keep pace with global advancements in AI technology [67]. Addressing these challenges demands concerted efforts to align legal frameworks with technological realities, ensuring that AI not only advances but does so in a manner consistent with the broader sociopolitical and ethical landscape.

Given the identified gaps in existing regulatory frameworks, Mexico’s approach to managing ethical concerns related to AI is multifaceted and strategic. Recognizing the global discourse on AI ethics, Mexico, along with other emerging countries like Jordan, recognizes the pressing need to address ethical considerations as a critical challenge in the AI landscape [67]. Mexico’s strategy includes not only aligning with international ethical standards but also adapting solutions to its specific socioeconomic context, thereby distinguishing it from wealthier countries that face similar ethical dilemmas [68,69].

Notably, Mexico was among the first nations to initiate discussions on AI regulation in 2018, emphasizing the need for each country to define its own AI regulatory framework, reflecting a proactive stance in shaping ethical guidelines tailored to its national priorities [3]. By integrating these considerations into its AI strategy, Mexico aims not only to protect against potential ethical pitfalls but also to leverage AI to foster local talent and innovation, thereby ensuring ethical AI development that is both locally relevant and globally informed [70]. This comprehensive approach underscores the need for ongoing dialogue and the adaptation of ethical frameworks to keep pace with technological advancements, highlighting the importance of collaborative efforts among nations to establish robust and context-aware AI policies.

When considering future legislative measures to improve AI regulation, it is crucial to adopt a comprehensive approach that encompasses both national and international perspectives. Nation-states recognize the need to establish national committees to ensure greater oversight of AI weaponry, which is aligned with broader national defense strategies [71]. However, as AI technologies transcend national borders, the need for international cooperation becomes evident.

Participation in international forums is essential to address domestic policy differences and align them with global regulatory standards. Such cooperation is particularly important when managing the strategic risks associated with lethal autonomous weapons systems (LAWSs), as unilateral actions by individual states are insufficient to address the complexities of these technologies [71]. Furthermore, policymakers are urged to collaborate with key allies to advance humanitarian objectives, emphasizing the ethical dimensions of AI regulation. This multilateral participation not only ensures a harmonized regulatory landscape but also fosters the creation of international standards that safeguard fundamental rights while encouraging responsible technological advancement.

It is important to note that the difficulty of regulating AI is tied to its characterization as artificial intelligence, with existing and new conflicting notions of the meaning of “artificial” and “intelligence” [72]. Therefore, future legislative measures must be adaptable, ensuring they nurture technological progress while remaining firmly anchored in ethical and social well-being.

5.1. Actionable Recommendations

In developing our recommendations, we considered the ten core principles that outline a human rights-centered approach to the Ethics of AI as defined by the United Nations Educational, Scientific, and Cultural Organization [73].

• Proportionality and Do No Harm. “The use of AI systems must not go beyond what is necessary to achieve a legitimate aim. Risk assessment should be used to prevent harm which may result from such uses”.
• Safety and Security. “Unwanted harms (safety risks) as well as vulnerabilities to attack (security risks) should be avoided and addressed by AI actors”.
• Right to Privacy and Data Protection. “Privacy must be protected and promoted throughout the AI lifecycle. Adequate data protection frameworks should also be established”.
• Multi-stakeholder and Adaptive Governance & Collaboration. “International law & national sovereignty must be respected in the use of data. Additionally, participation of diverse stakeholders is necessary for inclusive approaches to AI governance”.
• Responsibility and Accountability. “AI systems should be auditable and traceable. There should be oversight, impact assessment, audit, and due diligence mechanisms in place to avoid conflicts with human rights norms and threats to environmental wellbeing”.
• Transparency and Explainability. “The ethical deployment of AI systems depends on their transparency & explainability (T&E). The level of T&E should be appropriate to the context, as there may be tensions between T&E and other principles such as privacy, safety, and security”.
• Human Oversight and Determination. “Member States should ensure that AI systems do not displace ultimate human responsibility and accountability”.
• Sustainability. “AI technologies should be assessed against their impacts on ‘sustainability’, understood as a set of constantly evolving goals including those set out in the UN’s Sustainable Development Goals”.
• Awareness & Literacy. “Public understanding of AI and data should be promoted through open & accessible education, civic engagement, digital skills & AI ethics training, media & information literacy”.
• Fairness and Non-Discrimination. “AI actors should promote social justice, fairness, and non-discrimination while taking an inclusive approach to ensure AI’s benefits are accessible to all”.

In addition, according to the proposals of [74], we consider the following recommendations:

11.

The Judiciary must collaborate with AI, ethics, and human rights experts to develop ethical and legal guidelines and standards for the use and application of AI-based technologies in the judicial system.

12.

The capabilities of the National Electoral Institute (INE) should be strengthened through the creation of a regulatory framework that defines the limits and conditions under which AI-based tools can be used in political campaigns, electoral propaganda, and the dissemination of election-related information.

13.

The capabilities and regulatory framework of the National Institute for Transparency, Access to Information, and Personal Data Protection (INAI), as the competent body for personal data protection, should also be strengthened and modified to include the supervision of and compliance with regulations related to the use of AI-based technologies and personal data protection, and establish clear and specific regulations for data management in the cloud and in data centers.

14.

It is recommended to strengthen the capabilities of the Federal Economic Competition Commission (COFECE) as a regulatory body that promotes competition and equitable access to data so that new technologies, such as those based on artificial intelligence, are accessible to all.

It is also recommended to update the federal law on the protection of personal data held by private parties and the general law on the protection of personal data held by obligated subjects.

5.2. Possible Social Actors That May Influence Technological Decisions in the Country and Technology Regulation

Several social actors in Mexico influence the development of AI. In 2018, a national coalition of stakeholders from industry, academia, and civil society, called IA2030.mx, was launched to coordinate efforts, build a unified voice to address AI, and promote continued action and support for the topic. This coalition developed the first National AI Agenda, based on the collective intelligence of more than 400 people. However, the website (IA2030.mx) is no longer active.

According to the AI Readiness Index (https://oxfordinsights.com/ai-readiness/), accessed on 29 December 2025, Mexico does not currently play a leading role in Latin America and the Caribbean, ranking eighth. In fact, from 2020 to 2025, the country’s performance declined across key indicators (

Table 3. Mexico’s performance regarding AI, according to the AI Readiness Index.

Year	Position/Countries Evaluated	Overall Points	Position/Latin America
2020	55/172	49.36	5/32
2021	60/160	52.62	6/26
2022	62/181	51.78	7/31
2023	68/193	50.37	8/33
2024	71/188	53.29	8/32
2025	79/195	45.6	8/33

) and moved from fifth to eighth in the region.

To address this situation, we identified several social actors in Mexico who can contribute to AI regulation and the safe development of AI.

• The Center of Industrial Innovation and Artificial Intelligence (Centro de Innovación Industrial e Inteligencia Artificial, CII.IA).
• fAIrLAC+.
• The Mexican Academy of Computation (Academia Mexicana de Computación, AMEXCOMP).
• The Mexican Academy of Cidersecurity and Digital Law (Academia Mexicana de Ciberseguridad y Derecho Digital, AMCID).
• The Mexican Association of the Information Technology Industry (Asociación Mexicana de la Industria de Tecnologías de Información, AMITI).
• The Mexican Society of Artificial Intelligence (Sociedad Mexicana de Inteligencia Artificial, SMIA).
• The National Alliance for Artificial Intelligence (Alianza Nacional de Inteligencia Artificial, ANIA).
• The White Box Project.
• United Nations Development Program (PNUD Mexico).

Furthermore, we emphasize the importance of collaborating with key allies to advance humanitarian objectives, thereby ensuring that ethical considerations remain central to AI regulation. While Mexico’s approach to AI regulation is strategic and multifaceted, recognizing AI’s unique opportunities and challenges is essential to creating a robust legal framework that safeguards fundamental rights.

Ongoing dialogue in international forums highlights the need for cooperation to address policy differences and establish global regulatory standards that can guide nations in their AI governance. Despite these proactive measures, we recognize the limitations of traditional legal structures in addressing the rapid evolution of AI and the ethical dilemmas it presents. Therefore, future research should focus on developing innovative regulatory proposals that not only address current challenges but also anticipate the dynamic nature of AI technologies.

By fostering local talent and innovation, Mexico can leverage AI as an engine of social progress while ensuring that its development remains ethically sound and globally informed. In conclusion, the interplay between national and international regulatory measures underscores the need for an adaptive, inclusive legal strategy that positions Mexico favorably in the global AI landscape, ultimately contributing to the responsible advancement of AI technologies.