Dario Amodei
| Dario Amodei | |
| Occupation | Chief Executive Officer and Co-Founder |
|---|---|
| Employer | Anthropic |
| Website | https://darioamodei.com/ |
| Blog | https://darioamodei.com/ |
Dario Amodei is the co-founder and chief executive officer of Anthropic, an artificial-intelligence research company, and previously served as vice president of research at OpenAI, where he helped direct research that produced GPT-2 and GPT-3.[1]
Background[edit]
Stanford University[edit]
Amodei completed postdoctoral research at Stanford University School of Medicine after earning his doctorate at Princeton University. His work at Stanford involved applications of computational methods to biological research, including the analysis of cellular-protein networks. The postdoctoral work preceded his move into commercial artificial-intelligence research. [2]
His transition from computational biology to machine learning reflected a broader movement of researchers from academic science into industry laboratories during the expansion of large-scale machine learning. Amodei subsequently worked at Baidu and Google Brain, where he moved from biological and physical modeling toward neural-network research and speech recognition. [3]
Baidu[edit]
Amodei joined Baidu's Silicon Valley AI laboratory in 2014 and worked on speech recognition under Andrew Ng. He was a co-author of the research paper on Deep Speech 2, an end-to-end neural-network speech-recognition system that used large-scale computing and training data to recognize English and Mandarin speech. [4]
The Baidu work placed Amodei in an industry research environment in which advances in computing power, data and neural-network methods were changing the economics and technical possibilities of machine learning. Andrew Ng was a senior figure in the project, while researchers including Amodei, Adam Coates and Greg Diamos worked on the development of the speech system. [5]
Google[edit]
Amodei joined Google in 2015 as a senior research scientist on the Google Brain team. His work focused on deep learning and neural networks, and he was also involved in research on the safety of machine-learning systems. OpenAI identified his previous Google Brain work when it announced his hiring in 2016. [6]
At Google Brain, Amodei worked in a research environment that was increasingly centered on scaling neural-network systems with larger datasets and greater computing resources. His work on machine-learning safety also connected technical model development with questions about unintended behavior and reliability, themes that later became part of his responsibilities at OpenAI and the stated research focus of Anthropic. [7]
OpenAI[edit]
Amodei joined OpenAI in 2016 and progressed from work on AI safety to research leadership. He became research director and later vice president of research, with responsibility for setting research direction and leading teams working on large language models and longer-term AI safety. OpenAI said when announcing his departure in 2020 that he had collaborated with colleagues on GPT-2 and GPT-3 and had worked with Ilya Sutskever in setting the direction of the research organization. [8]
In 2019, when he was VP of research, Amodei was part of the OpenAI team that withheld GPT-2 citing safety risks. He chose not to release GPT-2's full weights because the model was "too dangerous" to release all at once. The AI community largely mocked the decision, calling it a PR stunt.[9]
Amodei's OpenAI tenure coincided with a shift in artificial-intelligence research toward increasingly large language models trained with extensive computational resources. His work included research into methods for incorporating human feedback into model training and the interpretability and safety of AI systems. At OpenAI, Ilya Sutskever was a key research collaborator and co-leader of the organization's research direction during Amodei's tenure. [10]
Anthropic[edit]
Amodei left OpenAI in December 2020 with a group of colleagues and co-founded Anthropic in 2021 with his sister, Daniela Amodei, and other former OpenAI employees. Anthropic was established as an AI research company focused on developing systems described by the company as reliable, interpretable and steerable. Its initial $124 million Series A financing was announced in May 2021, with Dario Amodei as chief executive officer and Daniela Amodei as president. [11]
As CEO, Amodei oversees Anthropic's research direction and strategic vision. The company developed the Claude family of AI systems and has expanded from a research-focused startup into a large commercial AI company whose products are used by businesses and other organizations. The company's status as a public benefit corporation places its stated public-benefit purpose alongside its commercial activities. [12]
Education[edit]
Amodei earned a bachelor's degree in physics from Stanford University and a doctorate in physics from Princeton University, where his doctoral work concerned neural circuits and electrophysiology. He was a Hertz Fellow during his graduate studies at Princeton. [13]
His doctoral thesis, Network-Scale Electrophysiology: Measuring and Understanding the Collective Behavior of Neural Circuits, received the Hertz Thesis Prize. His academic training combined physics, biophysics and computational neuroscience before his subsequent work in machine learning. [14]
Controversies[edit]
In 2026, Anthropic became involved in a dispute with the U.S. Department of Defense over restrictions on military uses of its AI systems. Amodei and Anthropic opposed removing safeguards that restricted autonomous weapons targeting and mass domestic surveillance, while Pentagon officials argued that government use should be governed by U.S. law rather than company-imposed restrictions. The dispute included negotiations between Amodei and Defense Secretary Pete Hegseth and a threat by the Pentagon to designate Anthropic a supply-chain risk. [15]
The dispute subsequently reached federal court. In September 2026, a U.S. appeals court upheld the Pentagon's designation of Anthropic as a national-security supply-chain risk after the company declined to remove the disputed safeguards. Anthropic argued that the designation was retaliatory and said it was considering further legal action; the appellate ruling followed a separate federal district-court decision that had rejected the government's designation on different legal grounds. [16]
References[edit]
- ↑ Leadership at Anthropic. Anthropic.
- ↑ Dario Amodei, PhD. Fannie and John Hertz Foundation.
- ↑ Fireside Chat with Dario Amodei. Microsoft Research.
- ↑ Deep Speech 2: End-to-End Speech Recognition in English and Mandarin. Proceedings of Machine Learning Research.
- ↑ Fireside Chat with Dario Amodei. Microsoft Research.
- ↑ Team update. OpenAI.
- ↑ Dario Amodei. Stanford Digital Economy Lab.
- ↑ Organizational update from OpenAI. OpenAI.
- ↑ Dario Amodei Warned Against GPT-2 in 2019. Now He's at the Centre of the Open-Source AI War.. Explainx.ai.
- ↑ Dario Amodei. World Economic Forum.
- ↑ Anthropic raises $124 million to build more reliable, general AI systems. Anthropic.
- ↑ Leadership at Anthropic. Anthropic.
- ↑ Dario Amodei, PhD. Fannie and John Hertz Foundation.
- ↑ Hertz Thesis Prize. Fannie and John Hertz Foundation.
- ↑ Anthropic digs in heels in dispute with Pentagon, source says. Reuters.
- ↑ US appeals court upholds Pentagon's blacklisting of Anthropic. Reuters.