iaS is the reference hub for AI safety in Spanish: we train and support professionals in their career development, and back projects dedicated to building safer AI.
We connect people working on AI safety across the Spanish-speaking world. Here you'll find local groups where you can meet people, learn and start contributing, in your own language.
The laboratories developing the world's most advanced AI are primarily located in the US, the UK, and China. Decisions about how it is deployed, who controls it, and what values it incorporates are, for the most part, made in english.
This is not just a technical detail: it is a safety challenge.
Safety evaluations of language models are designed in English, and their behavior varies across languages and risk categories. Spanish, an official language in 21 countries and present on every continent, lacks its own node to measure these differences, document them, and convey the evidence to regulators and developers.
iaS was born to be that node.

AI safety sessions in Spanish held since February 2026
people have registered for our sessions
people have returned to our sessions more than once
node specialized in AI safety in Spanish
the average rating attendees give us
Our speakers have worked or collaborated with
Apart Research
Google DeepMind
OpenAI
MIRI
Future of Life Institute
Center for AI Safety
LASR Labs
SaferAI
Pour Demain
Cambridge AI Safety Hub
ML4Good
Kairos
AI Risk Explorer
ORCG
PauseAI
AI Safety Barcelona
DLR
Max Planck
IIT
Princeton
TUM
UPM
UPC
UPV
UBA
Universidad de los Andes
UNAM
Universidad Politécnica de Yucatán
ITK Engineering
Mercado Libre
Sezzle
TShaped
Come with a question, a project in progress, or simply curiosity. Follow our Luma calendar to take part in our activities.

Specialized program on Artificial General Intelligence (AGI) strategy, based on the BlueDot Impact curriculum.
| Duration | 5 weeks (25 h in total) |
|---|---|
| Sessions | 2 h online per week for 5 weeks |
| Self-study | about 3 h per week of readings and exercises |
| Format | Online |
| Certification | Yes, iaS certificate based on the BlueDot Impact curriculum |
| Selection | By application, 4 to 10 people per cohort |

Why would an AI that does not understand Spanish be a systemic risk? Analysis of safety failures in Spanish-language language models.

Together with BlueDot Impact, we run AI safety seminars, a space to connect with experts from around the world and have essential conversations about the future of technology in our language.
| Duration | 1 h |
|---|---|
| Frequency | First and third Wednesday of each month |
| Workload | No preparation needed, just a desire to learn |
| Format | Online |
| Certification | No certificate |
| Selection | Open registration |

Join iaS's Discord and discuss AI safety in Spanish with other people; this is where ideas and collaborations are born, and the next one could come from your conversation.