Dario Amodei
BLUF
Dario Amodei is CEO and co-founder of Anthropic, which he established in 2021 after departing as VP of Research at OpenAI with a cohort of safety-focused researchers. A physicist by training (PhD, UC San Diego), he has been the primary public voice for Anthropic’s policy positions, including the Responsible Scaling Policy (RSP), the Constitutional AI alignment framework, and the Project Glasswing consortium that provides frontier Claude Mythos access to cybersecurity organizations. His April 2026 meetings at the White House — at the same time as NSA operational testing of Claude Mythos — were reported by Axios as attempts to navigate the administration’s opposition to expanding the Glasswing consortium.
Key Facts
- Role: CEO and co-founder, Anthropic (2021–present)
- Previous: VP of Research, OpenAI (2016–2021); lead on GPT-2/GPT-3 development
- Sibling: Daniela Amodei (President and co-founder, Anthropic)
- Education: BS Physics, Stanford; PhD Physics, UC San Diego
- Public positions: Advocates for frontier AI safety research as the necessary precondition for frontier AI capability development; testified before US Senate on AI risk (2023)
- Glasswing/Mythos: Personally involved in the White House negotiations over Glasswing expansion; met with officials in late April 2026 to address administration concerns (Axios, 2026-05-01)
Analytical Significance
Amodei represents the “safety-through-capability” school of AI development: Anthropic’s argument is that safety-conscious frontier labs must lead capability development, or capability development will be led by less safety-conscious actors. This framing justifies continued frontier model development (Mythos) despite dual-use risks.
His engagement with US government — White House, IC, DoD — positions Anthropic and himself at the center of the US AI governance debate. The contradictions in that space (NSA using Mythos while DoD designates Anthropic a supply-chain risk) affect Anthropic’s regulatory and commercial trajectory directly.
Key Connections
- Anthropic — CEO and co-founder
- Claude Mythos — the Mythos/Glasswing model is Anthropic’s most advanced frontier product; Dario Amodei is its principal public advocate
- Constitutional AI — alignment framework Amodei co-developed
- Responsible Scaling Policy — Anthropic’s tiered capability-risk framework; Amodei is the lead signatory
Sources
- Axios, 2026-04-29 and 2026-05-01: White House meetings and Glasswing reporting — [Medium-high confidence]
- Anthropic official bio — [High confidence — primary]
- US Senate testimony (2023): AI risk — [High confidence]