AI Safety Barcelona connects students, researchers and institutions across Barcelona and Spain to the global work of making frontier AI safe.
We can measure what a model outputs. We still cannot say reliably what it is computing internally, or whether it will keep doing what we asked as it becomes more capable.
Hiring, credit, medical triage, public administration. Those systems run in Spain today, almost all bought from abroad and deployed by institutions with no capacity to audit them.
The AI Act is being implemented across Europe. Most of that work happens in Brussels, London and San Francisco. Barcelona's universities produce exactly the people who could do it.
Workshops that build technical and governance context, alongside reading clubs and paper presentations. No prerequisites.
Open events in Barcelona, and introductions to the people already working on this — here, and in the labs and institutions abroad.
Hackathons: a weekend to go from reading about AI safety to producing something of your own, with mentors on hand.