Redwood Research
Berkeley nonprofit researching AI control and risks from misaligned AI.
At a glance
People
Buck Shlegeris (CEO), Ryan Greenblatt (Chief scientist) and 2 others
See all 4Connections
Also on this page: Key facts · Links · Sources · Related pages
- Redwood Research is a 501(c)(3) nonprofit in Berkeley that studies risks from powerful AI and how to reduce them.
- It introduced the research area of AI control, with protocols for monitoring LLM agents that may be working against their users.
- With Anthropic it showed in Alignment Faking in Large Language Models that Claude sometimes hides misaligned intentions.
- It advises AI companies including Google DeepMind and Anthropic.
Key facts
- Kind
- 501(c)(3) nonprofit, Berkeley1
- Focus
- AI control; evaluations of strategic deception; consulting on misalignment risk1
People
Led by
- Buck Shlegeris
CEO 2
- Ryan GreenblattPerson
Chief scientist 3
Board
Buck Shlegeris, Nate Thomas, Ammon Bartram 2
Connections
Partners
- AnthropicOrganization
co-wrote Alignment Faking in Large Language Models 1
Links
Sources
Last checked against its sources today
Is something on this page wrong or out of date? Tell us
How recent
The line under a date fills as time passes: empty when new, then one equal share each for a day, a week, a month and a year. For episodes and other items it is full at 2 years; for "last checked", at a year. A striped line means the date is known only to the month or year.