Grantmaking.ai project link

What this project is

BAISH is the largest AI Safety hub in Latin-America. Founded in 2024, we have a WhatsApp community with over 250 members. As a whole, we've been instrumental in placing at least 5 of our members (including directors) into full-time AI Safety roles, working at companies such as Equistamp, 80000 hours, and Apart Research. Furthermore, Our activities include, but are not limited to:

- Fellowship Program: Supported by CG, our community started the AISAR program; an in-person fellowship for technical AI Safety research. The first edition ran through the 2nd semester of 2025 for 6 scholars, and will now run again for 15 scholars.

- Courses: Supported by 3 rapid grants from BlueDot, we have run 8 BlueDot cohorts, and in the past we've run 4 other courses similar to the old AI Safety Fundamentals course, and a research workshop. We've had a total number of over 80 people

- Community Events: We run monthly networking and social events, which usually have an in-person attendance of 50-80 people, with an average of ~50% of them being newcomers. We have a Luma website with almost 500 subscribers. We've also supported part of a 16-person Argentinian delegation going to the last EAG London (May 2026).

- Communication: We have a dedicated team of volunteers and paid community managers in charge of our communication. This mainly includes the instagram profile.

Team

Founding team and Directors:

- Eitán Sprejer

- Research Manager Lead at Apart Research

- Carlos Giudice

- Research Engineer at Equistamp

- Luca De Leo

- Product Engineer at 80,000 Hours

Advising team:

- Pablo Stafforini

- AI Enablement Manager at Epoch

- Sergio Abriola

- Professor at University of Buenos Aires

- Iván Arcuschin

- Lead Research Scientist at Poseidon Research

- Agustín Martinez Suñé

- Oxford Post-doc working on Formal Verification x AI Safety

- Tobías Martin

- MIRI

- Tao Burga

- Technology Fellow at the Institute for Progress

- Luis Enrique Urtubey:

- Director of CEGIA

Managers and leadership:

- Nicolás Spinelli

- Marina Romanisio

- Guido Bergman

- Rocío Monjes

- Nicolás Martorell

Team and Volunteers:

- Gonzalo Heredia

- Gaspar Labastie

- Tobías Bersia

- Santiago Nuñez Rimedio

- Augusto Masciotta

- Juan Charovsky

- Trinidad Borrell

- María Borrell

Currently, BAISH is being run by a (growing) team of 16 volunteers, who can each put a limited number of hours into the project. The funding we've received so far only manages to cover for regular day-to-day spendings, and cannot (contractually) be used to pay for salaries. We've recently hired a part-time ops lead, as well as some community managers, all paid out of pocket with the money the directors donated. In order to grow, our current limitation is both seniority and time-commitment. In order to secure senior, full-time people, we concluded we need to have our own source of funding for paying salaries.

What's your track record on similar projects?

We've been instrumental in placing 5 of our members (including directors) into full-time AI Safety roles, working at companies such as Equistamp, 80000 hours, and Apart Research. Additionally, we've helped get 20 of our members placed into external AI safety fellowships, and upskilling programs, including MARS (4), SPAR (4), ML4Good (8), ARENA (2), Cambria (2), CLR, Algoverse, CORDA, Apart Secure Program Synthesis Fellowship, and Iliad.

Theory of impact

The AI Safety field building efforts are not on track to meet the demand that the challenge presents. We need more people working in making AI go well, decreasing extinction level risks. Most of the talent is currently not working on this most important problem, and there are diverse sources of talent distributed throughout the world. Buenos Aires, Argentina (16M people) has one of the top-100 universities in the world, as well as a moderately big startup and tech ecosystem. We aim to reduce x-risks by tapping into the Argentinian talent pool,  help them go through the necessary upskilling and get them ready for placement in impactful positions, be it locally or abroad.

What are the most likely causes and outcomes if this project fails?

If the project fails, the most likely cause would be not finding the right person for the roles. We currently have one person in mind who we think has a great fit for the position.

How much money have you raised in the last 12 months, and from where?

grantmaking.ai**** Launch Round grant

from grantmaking.ai

$27,000

Kairos - Pathfinder Fellowship

from Kairos

$6,000

BlueDot Rapid Grants for In-person AIS Courses

from BlueDot Impact

$15,000

Coefficient Giving - University Group Organizer Fund

from Coefficient Giving

$12,500

How the money will be spent

27k: 1.5 FTEs for senior roles in Buenos Aires: 3k / month x 6 months. This would cover for a full-time community lead (someone that directly does talent-scouting, community events, connections with other organizations, plus funding for a part-time ops lead.

45k: 2.5 FTEs for senior roles in Buenos Aires: 3k / month x 6 months. This would cover for all of the above, plus someone that could lead the governance initiatives we have planned. We're planning on developing a governance think-tank that aims to work on advising and targeted outreach to our local politicians, plus our own technical governance research. We have some great advisors for this, like Luis Enrique Urtubey, and Tao Burga.