Ben Mann is a co-founder of Anthropic. Mann's own site says Mann spent six years as a software engineer at Google, then worked at OpenAI and MIRI before co-founding Anthropic, and that as of 2023 Mann was leading product engineering there. Anthropic's January 2026 announcement of its expanded Labs team, which incubates experimental products at the frontier of Claude's capabilities, said Mike Krieger would join Labs to build alongside Mann, and in October 2026 Anthropic named Mann as the co-founder who signed its memorandum of understanding with Singapore's Ministry of Digital Development and Information. Mann is the second-listed author of "Language Models are Few-Shot Learners" (2020), the GPT-3 paper, and a co-author of Anthropic's 2022 papers on training a helpful and harmless assistant with RLHF and on Constitutional AI.123456
At a glance
- Current role
- Co-founder
- Company
- Anthropic
- Sector
- Company
- Work listed
- 3
- Sources cited
- 7
- Profile reviewed
Career
Ben Mann's personal site describes six years as a software engineer at Google, followed by starting a company and then turning to AI safety through work at OpenAI and MIRI. Mann then co-founded Anthropic and, as of 2023, was leading product engineering there.1
Anthropic's January 2026 announcement of an expanded Labs team, focused on incubating experimental products at the edge of Claude's capabilities, said Mike Krieger would join Labs to build alongside Mann. Mann has also represented Anthropic as a co-founder at public events, sharing the opening keynote with Dario Amodei at the Tokyo Builder Summit in October 2025.27
In October 2026 Mann signed a memorandum of understanding between Anthropic and Singapore's Ministry of Digital Development and Information, covering technical exchanges on model evaluation, testing standards and incident response, as Anthropic opened an office in Singapore.3
Research
Mann is the second-listed author, after Tom B. Brown, of "Language Models are Few-Shot Learners" (2020), the paper that introduced GPT-3. At Anthropic, Mann co-authored the 2022 paper on training a helpful and harmless assistant with reinforcement learning from human feedback, and the December 2022 Constitutional AI paper on training harmless assistants from AI feedback guided by a set of principles.456
Work and contributions
Language Models are Few-Shot Learners (GPT-3) Published research
Second-listed author, after Tom B. Brown, of the 2020 paper that introduced GPT-3.4
Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback Published research
Co-author of the 2022 Anthropic paper on applying preference modeling and RLHF to fine-tune language models as helpful and harmless assistants.5
Constitutional AI: Harmlessness from AI Feedback Published research
Co-author of the 2022 Constitutional AI paper.6
Timeline
- Second-listed author of "Language Models are Few-Shot Learners", the GPT-3 paper.4
- Co-authored Anthropic's paper on training a helpful and harmless assistant with RLHF.5
- Co-authored "Constitutional AI: Harmlessness from AI Feedback".6
- Leading product engineering at Anthropic, according to Mann's personal site.1
- Gave the opening keynote with Dario Amodei at Anthropic's Tokyo Builder Summit.7
- Named in Anthropic's announcement of the expanded Labs team, which Mike Krieger joined to build alongside Mann.2
- Signed Anthropic's memorandum of understanding with Singapore's Ministry of Digital Development and Information.3
Other documented roles
Roles the sources document that are not confirmed as current.
Where they work
- Anthropic 34 people in the wiki
Related people
- Evan HubingerAlso at Anthropic
- Ethan PerezAlso at Anthropic
- Samuel MarksAlso at Anthropic
- Samuel R. BowmanAlso at Anthropic
- Amanda AskellAlso at Anthropic
- Wojciech ZarembaSafety and alignment · Language models
- Aylin CaliskanSafety and alignment · Language models
- Shengjia ZhaoLanguage models · Safety and alignment
- Margaret MitchellSafety and alignment · Language models
- Ilya SutskeverLanguage models · Safety and alignment
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
7 sources · profile reviewed Oct 21, 2026 · data updated Oct 21, 2026
- Ben Mann — personal site (bio section)
- Introducing Labs
- Anthropic opens Singapore office and signs MOU with the Government of Singapore
- Language Models are Few-Shot Learners (arXiv 2005.14165)
- Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback (arXiv 2204.05862)
- Constitutional AI: Harmlessness from AI Feedback (arXiv 2212.08073)
- Tokyo Builder Summit (speakers list Ben Mann as co-founder, Anthropic)