The largest AI safety hub in Latin America
In-person courses, a research lab and a community of 400+ in Buenos Aires, for anyone who wants to help AI go well.
About BAISH
If AI goes well, it could help with problems that have been stuck for decades. If it doesn't, the harm could be very hard to reverse. Nobody knows how it will go, and some researchers we respect aren't convinced there's a serious risk at all. We think it is plausible enough, and the work neglected enough, that the people working on it can change the outcome.
The case for taking AI seriouslyArgentina has the talent, the universities and the research culture. What was missing was a place where people could get into AI safety, and keep doing it, without leaving the country.
Founded in 2024. Run by volunteers and a small paid staff, alongside our sister fellowship, AISAR.
What we do
Technical AI Safety, the Technical AI Safety Project and Frontier AI Governance, taught in person in Buenos Aires.
Monthly socials, an annual retreat, hackathons and a WhatsApp group where members share opportunities and papers.
Our research arm: a Buenos Aires-based group working on interpretability and control, alongside the AISAR fellowship.
One-on-one career support, help preparing applications and trips to conferences like EAG.
Programs
Free in-person courses run with BlueDot Impact, and a research fellowship with a stipend. Courses select participants through a short application.
BlueDot Impact’s technical AI safety course, hosted in person by BAISH. Build the foundations to understand current safety techniques and where you can contribute.
A project course for making a first concrete contribution to AI safety research or engineering, hosted in person by BAISH.
Explore governance challenges from frontier AI systems: coordination, compute governance, accountability, and risk assessment.
Six-month mentored research fellowship in Buenos Aires, funded by Coefficient Giving.
Tell us a bit about yourself through the contact form and we'll point you to the right one.
Members
BAISH members have moved into full-time AI safety roles, international fellowships and published research.
See all stories“Thanks to the group and to Eitan, who backed me a lot through the application process.”
MIRI
ApplyathonMARS
“BAISH gave me the foundation and connections to transition into AI safety research professionally.”
Former Courses Lead · Research Engineer, Equistamp
Courses LeadAlgoverseEquistamp
Took the intensive course in March 2026 and was accepted to ARENA two months later, where his team's project won first place. He now leads research at BAISH Labs and runs the Technical AI Safety Project.
Technical AI Safety Project lead · Research Lead, BAISH Labs
TAIS IntensiveARENABAISH Labs
Community
About half the people at our monthly socials are newcomers.
Every month in Buenos Aires, typically with 50–80 people.
400+ members sharing opportunities, papers and events.
Our activities, events and opportunities, by email.
We talk to journalists and work with universities, companies and other communities.
