Building Hispanic participation in advanced AI safety and helping Hispanic-serving institutions advocate for effective AI governance.
Endorsements made here support Hispanic AI Safety Institute.
Building Hispanic participation in advanced AI safety and helping Hispanic-serving institutions advocate for effective AI governance.
Endorsements made here support Hispanic AI Safety Institute.
We are launching the first AI Safety organization focused only on educating and advocating for hispanic communities in the U.S.
Hispanic residents account for 21% of the U.S. population, approximately 70.1 million people in 2025. As governments decide how powerful AI systems should be developed and overseen, HASI aims to help the institutions serving these communities participate with the knowledge and capacity to advocate effectively.
We will develop research briefs and educational materials in Spanish, then work with journalists and Hispanic-serving organizations to put them to use. Workshops and continuing relationships will connect explanations of advanced AI risks to concrete policy questions. We intend to support informed consultation responses and dialogue with policymakers, complementing existing Latino technology-policy advocacy with dedicated attention to catastrophic risks and human control.
HASI is led by co-founders Tobias Martin and Debora. Tobias works in AI safety communications at MIRI and holds a master’s degree in Learning Design, Innovation, and Technology from Harvard. Debora is a PhD student. HASI is at the launch stage: draft educational materials and research syntheses exist, while the institutional program and delivery team are being developed.
We are raising for 12 months of operations. Funding would provide dedicated staff time and the editorial, research and operational support needed to deliver the program. We will start with institutional pilots and expand based on demonstrated use.
Effective safeguards for advanced AI require public institutions capable of understanding the risks and holding developers accountable. Hispanic-serving organizations could contribute through their community relationships and policy engagement. HASI will help interested institutions develop the specialist knowledge needed to participate substantively.
Our intended pathway begins with bilingual explanations and sustained work with institutional partners. As partners develop a better understanding of dangerous capabilities and failures of human control, they can make more informed policy contributions. HASI will support that engagement with research briefs and analysis of proposals for independent evaluation and oversight.
If this work helps strengthen the adoption or implementation of effective safeguards, it could reduce the likelihood that dangerously capable systems are deployed without adequate evaluation or control.
The additional value we hope to provide is engagement by institutions that would otherwise devote limited attention to advanced AI safety, alongside specialist support for existing advocates. We will test this through repeat institutional use, assessments of understanding and documented examples of partners applying the work in reporting or policy engagement.
Team Member
No comments yet. Be the first to share your thoughts.