Yuntao Bai is a researcher at Anthropic. Google Scholar lists the affiliation Anthropic, PBC, with a verified anthropic.com email. Bai is first author of "Constitutional AI: Harmlessness from AI Feedback" (2022) and of "Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback" (2022).123
At a glance
- Current role
- Researcher
- Company
- Anthropic
- Work listed
- 1
- Sources cited
- 3
- Profile reviewed
Accomplishments
Constitutional AI
The Constitutional AI paper's author-contributions section says Bai developed the method further and designed and carried out most of the experiments, and that Bai and Jared Kaplan drafted the paper. Anthropic's December 2022 post summarizes the method: a list of principles (a "constitution") plus self-critique, revision, and RL from AI feedback.12
Work and contributions
Constitutional AI: Harmlessness from AI Feedback Published research
First author of Anthropic's Constitutional AI paper on training a harmless assistant from a written list of principles.12
Where they work
- Anthropic 49 people in the wiki
Open roles at Anthropic (675)
- Data EngineerSan Francisco, CA | New York City, NY · Seattle, WA · On-sitePosted Oct 30, 2026
- Engineering Manager, Integrity and AttestationSan Francisco, CA | Seattle, WA · On-sitePosted Oct 30, 2026
- Contracts Manager, GTMSan Francisco, CA · On-sitePosted Oct 30, 2026
Related people
- Samuel MarksAlso at Anthropic
- Collin BurnsAlso at Anthropic
- Ben MannAlso at Anthropic
- Arthur ConmyAlso at Anthropic
- Evan HubingerAlso at Anthropic
- Jonas AndrulisLanguage models · Safety and alignment
- Kenneth LiLanguage models · Safety and alignment
- Shengjia ZhaoLanguage models · Safety and alignment
- Zvi MowshowitzLanguage models · Safety and alignment
- Margaret MitchellSafety and alignment · Language models
Suggest a correction or removal
Profiles cover public professional work only, and each statement cites a source. If something is out of date, missing or wrong, tell us and link a source. People can also ask for their profile to be corrected or removed.
References
3 sources · profile reviewed Oct 25, 2026 · data updated Oct 25, 2026
AI People Wiki is free and open source. Support the wiki — buy me a coffee