Our charter

A meaningful share of philanthropic money earmarked for "AI risk" never reaches organizations doing direct, honest work on existential risk from artificial superintelligence (ASI). It is diverted toward adjacent causes — bias, deepfakes, job displacement — or ends up sustaining conferences and personal networks rather than public-facing work. CASIR exists to close that gap: to get money to organizations and projects that state plainly, to the public and to power, that ASI could cause human extinction or irreversible loss of human control — and that carry themselves like it is a serious, adult subject rather than a subculture. CASIR's purpose is to raise the field's sanity waterline — general comprehension of the risk — not to build a movement, a brand, or a leadership class. Grants favor clarity reaching many people over influence concentrated in a few hands.

Pillars

1. Pillar I — Clarity (Scope). Funds go only to work squarely about existential risk from advanced or superintelligent AI: extinction risk, loss of human control over ASI-level systems, irreversible human disempowerment. Algorithmic bias, deepfakes and misinformation, job displacement, or "AI ethics" broadly, are out of scope as a primary framing. CASIR Members (e.g. PauseAI, ControlAI) have ASI/x-risk as their stated core mission and receive general operating support; CASIR Partners (e.g. CAIS) need not be central to the mission but regularly produce work addressing it, and are funded per-project only.
2. Pillar II — Straightforwardness (Method). Funded work engages directly — with the public, policymakers, and elites — in plain, concrete language a non-specialist can follow: op-eds, testimony, public talks, direct policymaker briefings, plain-language explainers. Multi-step "theories of change" that substitute for direct communication, or strategies premised on concealment, flattery-for-access, or indirection, are disqualifying.
3. Pillar III — Professionalism (Conduct). Funded organizations, and the public-facing staff who speak in their name, keep personal or sexual life off their public/professional platform, do not manufacture scandal or controversy for attention, and do not publicly insult named individuals or make Nazi/Godwin comparisons. Grant agreements include a conduct clause; a violation triggers review and can result in clawback of unspent funds.