Dario Amodei

Dario Amodei trained as a scientist before becoming an AI researcher and company founder. He completed his undergraduate studies at Stanford University and earned a doctorate at Princeton University, where his work was in biophysics. His path into machine learning connected an interest in complex scientific systems with the growing capabilities of neural networks.

Amodei worked at Baidu and then OpenAI, where he became vice president of research. His work there included involvement with the GPT-2 and GPT-3 language-model programs. He also co-authored Concrete Problems in AI Safety, a 2016 paper that organized practical research questions such as unintended behavior, reward exploitation, and robustness outside training conditions.

In 2021, he co-founded Anthropic with colleagues including his sister Daniela Amodei. As CEO, he leads a company whose public research agenda places substantial emphasis on the reliability, interpretability, and safety of advanced models. Anthropic develops Claude, a family of AI assistants and models offered through consumer interfaces and developer services.

His career combines technical research with the institutional decisions required to build and deploy frontier AI. That includes choosing research priorities, developing evaluation practices, and directing a business around large-scale model development. His earlier scientific training and safety research provide context for Anthropic’s stated focus on understanding how increasingly capable systems behave.