AI Safety
Hub_
A physical institution for the age of advanced AI.
Built for long-term thinking, interdisciplinary collaboration, and civic legitimacy.
A place where researchers, engineers, policymakers, artists, founders, and the public work through the implications of AI together.
A space designed to support:
- AI safety research
- Interdisciplinary collaboration
- Civic dialogue
- Artistic and cultural exploration
- Public understanding of AI
- Long-term coordination
Where technical work can stay connected to society, culture, governance, and everyday human life.
Designed for
permeability.
The architecture mirrors the culture. Public life at the base, shared work in the middle, quiet research above. Ideas move up, down, and across.
Research residencies, deep work, long-term projects, and focused collaboration.
Coworking, salons, workshops, studios, and interdisciplinary events.
Cafés, exhibitions, lectures, dining halls, libraries, and public programs.
A third place for the AI transition
Third places were pivotal in the periods of social and cultural changes of the past. Cafés in Enlightenment Europe. Research labs like Bell Labs. Artistic salons, universities, hacker spaces, monasteries, libraries. Places where people from different worlds met one another repeatedly over time, and where ideas matured through conversation, trust, criticism, and collaboration.
AI Safety Hubs are enduring places for research, collaboration, public learning, cultural exploration, and long-term thinking. Holding containers for conversations that won’t fit cleanly inside existing institutions.
Advanced AI will reshape science, labor, governance, media, education, and daily life. The transition deserves physical institutions designed for reflection, responsibility, and collaboration across disciplines.
Four pillars, one institution.
Technical AI safety, governance, forecasting, interpretability, alignment, and long-term risk work, informed by perspectives from outside the lab.
Applied systems, infrastructure, prototypes, and experiments, grounded in responsibility and long-term thinking.
Workshops, fellowships, lectures, and public programs that help people understand the challenges and implications of advanced AI.
Art, philosophy, exhibitions, reading groups, writing, and civic dialogue around the societal implications of AI.
Advanced AI touches every layer of life: social, technical, personal. No single discipline can hold all of it. An institution can.
A network of local institutions.
Each Hub belongs to its city. Built by local people, shaped by local culture, responsive to local realities. The network lets perspectives circulate between cities without flattening them into a single worldview.
Singapore
Pragmatic, infrastructural, deployment-oriented. A hub focused on governance, coordination, and AI in real-world systems.
Boston
Academic, technical, biosecurity-oriented. Closely tied to universities, labs, and frontier research communities.
London
Theoretical, cultural, reflective. A hub where philosophy, governance, and art meet frontier AI questions.
The institutions built during periods of technological transition shape society for decades afterward.
Advanced AI is arriving faster than the institutions needed to meaningfully integrate it into public life.
If spaces for AI discourse are not built intentionally, they will emerge accidentally: inside corporations, inside governments, inside fragmented online ecosystems.
The institutions shaping the AI era are already forming. AI Safety Hub is proposed intentionally, as an institution built for long-term thinking, public engagement, and interdisciplinary collaboration.
Toward our first physical location.
We are looking for people interested in helping shape a new kind of civic and intellectual infrastructure for the AI era, as well as funders and in-kind donors from governments, companies and philanthropic organizations.
- Researchers
- Engineers
- Designers
- Artists
- Founders
- Policy thinkers
- Educators
- Institution builders