2021
Dario Amodei
b. 1983
Safety & Alignment
Co-founded Anthropic to build capable models like Claude with an emphasis on safety and constitutional training.
Formerly a research leader at OpenAI and co-author of work on scaling laws, Amodei co-founded Anthropic in 2021.
Anthropic's Claude models are trained with 'Constitutional AI', using written principles to guide behaviour.
Dated contributions
2021
Safety & Alignment
Dario Amodei
Co-founded Anthropic to build capable models like Claude with an emphasis on safety and constitutional training.
2022
Constitutional AI
Yuntao Bai & the Anthropic research team
Anthropic researchers explore training assistants with written principles and AI feedback.
Paper · 2022 — Constitutional AI: Harmlessness from AI Feedback ↗
