A nonprofit organization that brings together AI research, education, and governance from Mexico.
Mission
AI Safety Mexico is a nonprofit organization founded in 2025 that advances research, education, and responsible governance of artificial intelligence in Mexico. We work at the intersection of three communities — researchers, educators, and policymakers — to ensure that AI systems are developed and deployed safely, aligned with human values, and attentive to the Mexican context.
Our work combines open research, context-aware training, and collaboration with partner institutions. We operate as a multidisciplinary team and partner with national and international organizations that share our commitment to giving the Global South a voice of its own in the conversation about AI safety.
Vision
We envision a Mexico where the Global South speaks with a voice of its own in the technical and normative definition of AI safety: a consolidated community of researchers, educators, and policymakers producing locally grounded knowledge, training new generations, and taking an active role in shaping regulatory frameworks attentive to Mexican realities.
A Mexico where the Global South speaks with a voice of its own in the technical and normative definition of AI safety.
Strategy
We are not building a community to “join.” We are building a launchpad toward the frontier of AI safety and Mexico's credible research voice: a handful of high-agency Mexicans pursuing the frontier, a citable body of research, and a Spanish-language bridge to the people who will decide how Mexico faces advanced AI.
Seed in Mexico a body of work that produces original research in key areas of AI safety. Our reference point is the AI Safety Institutes.
Two effects follow from that: acting as a Spanish-language translation layer for frontier AI, and as a credible, nonpartisan voice that a Mexican decision-maker can call.
That is our long-term direction; everything we do today builds toward it.
Help a small number of high-agency people propel themselves toward the frontier, produce a citable body of research, and translate it for Mexican decision-makers and for the Spanish-speaking world.
All of it building on infrastructure that already exists (Apart, SPAR, BlueDot, Atlas, academic institutions in Mexico) rather than building our own from scratch.
Four axes
The areas where we concentrate our efforts. They are not watertight compartments: most projects cut across several of them.
We produce original and collaborative work in alignment, mechanistic interpretability, and technical governance, in partnership with national and international institutions.
We design and deliver open and bespoke courses — for communities, universities, and research centers — that adapt contemporary AI safety frameworks to the Mexican context.
We produce technical analyses, take part in working groups, and publish specialist documents that feed into public consultations and AI policy design in Mexico.
We sustain ongoing spaces for dialogue — working groups, workshops, conferences — that connect those who research, develop, and regulate AI in Mexico.
Origin
AI Safety Mexico was born in 2025 in response to an observation shared by a group of researchers, engineers, and practitioners in Mexico: the conversation about AI safety is moving at a fast pace in the Global North, while the concerns, values, and contexts of the Global South are rarely brought into that conversation.
We decided to organize ourselves to close that gap from a specific country — Mexico — and through a specific practice: rigorous research, open training, and the production of technical inputs that support informed decision-making about AI. The organization was formally established in 2025 with a multidisciplinary founding team combining experience in machine learning, social sciences, public policy, and applied mathematics.
Since then, we've run open and institutional courses, published in national venues, taken part in international events (EAGxCDMX, ML4Good, SPAR), and partnered with allies such as Centro Geo, UADY, ENAIS, and Apart Research. Every activity follows the same principle: producing locally grounded knowledge that is useful to those who research, regulate, or deploy AI in Mexico.
Identity
The minimum worth knowing about us before exploring the rest of the site.
Behind AI Safety Mexico is a multidisciplinary group of researchers, engineers, and educators committed to AI safety in Mexico.
Meet the team