FAR.AI
- Type
- Civil society organisation
- Lieu
- États-Unis — United States (Berkeley)
- Dernière vérification
- 2026-09-22
- Prochaine vérification
- 2027-03-21
Pas encore traduit — affiché en anglais.
Comment les joindre
- Contact page
- Grants for academics and independent researchers
- FAR.Labs Berkeley co-working space
Ce canal est conservé uniquement comme référence. Sa vérification est absente, date de plus de 180 jours, ou la période de contact n'est pas ouverte actuellement ; ne vous y fiez pas avant une nouvelle vérification.
- Careers
- Newsletter
- Homepage
Ce qu'il fait
Runs three distinct arms: in-house research on a portfolio of agendas (robustness, deception detection, red-teaming, safeguard vulnerabilities); a field-building events programme (Alignment Workshops, the Seoul and Cambridge workshops, verification workshops, technical-governance convenings); and FAR.Labs, a Berkeley co-working space for AI safety researchers, plus a grantmaking programme for academics and independent researchers.
Évaluation franche
Genuinely a connector organisation and a good place for a technically credible outsider to get into the room via workshops or FAR.Labs. Weakness for routing purposes: no published email, only a form.
Préoccupations couvertes
Broad frontier-risk portfolio — reward hacking, deception, jailbreaking through layered defences, and misuse. Deliberately agenda-agnostic; its comparative advantage is convening rather than a single thesis.
Canal gouvernemental
Indirect but useful — its workshops convene government, lab and academic participants, including around the safety-summit circuit.