We produce open research on technical AI safety, governance, and policy questions specific to Mexico, in collaboration with national and international institutions.
Lines of work
Four fronts where we focus our questions, methods, and publications.
Alignment, mechanistic interpretability, and red-teaming applied to the study of deployed models.
Technical analysis and policy instruments for the responsible development and deployment of AI.
Mexico-based studies on perceptions, risks, and priorities of the Global South.
Materials, curricula, and open publications that connect research with training in Mexico.
Publications
Results that the team and its collaborators release in open venues.
Trabajo destinado a capturar preocupaciones locales acerca de la seguridad en inteligencia artificial en Mérida, Yucatán. La encuesta piloto explora percepciones, riesgos y prioridades locales como insumo para la agenda de AI safety en México.
Ongoing
Active team projects and international collaborations.
A platform and paper that map the international AI safety community through taxonomy and semantic search. Output of the SPAR program with an international team.
See the projectCollaborations
Partner organizations that sustain and enrich our research agenda.
Have a research question, a dataset, or a collaboration in mind? Let's connect to explore how we could work together.
Propose a collaboration