Carroll L. Wainwright
Also known as Carroll Wainwright
Carroll L. Wainwright worked on AI safety, from reinforcement-learning benchmarks to the training of language models. Together with Peter Eckersley, he released SafeLife, a reinforcement learning environment for testing agent safety, in December 2019. In 2022 he co-authored Training language models to follow instructions with human feedback. In June 2024, he left OpenAI, helped organize the public Right to Warn letter, and signed it.
Timeline
-
December 2019
Releasing SafeLife
Together with Peter Eckersley, he releases SafeLife, a reinforcement learning environment for testing agent safety, in December 2019.
-
2020
An updated SafeLife paper
An updated version of the SafeLife 1.0 paper is presented at the AAAI SafeAI 2020 workshop.
-
October 2020
A SafeLife leaderboard
In October 2020 he introduces a SafeLife leaderboard, a competitive benchmark for safer AI.
-
2022
Training models with human feedback
Training language models to follow instructions with human feedback, a 2022 paper, lists him as a co-author.
-
June 2024
Leaving OpenAI
In early June 2024 he leaves OpenAI, the same week he helps organize the public Right to Warn letter. The letter's signatory list names him as formerly of OpenAI.
Sources and credits
Photo: We have no published photo of Carroll L. Wainwright yet. Send us a free-licensed photo.
- Introducing SafeLife: Safety Benchmarks for Reinforcement Learning
- SafeLife 1.0: Exploring Side Effects in Complex Environments
- Introducing the SafeLife Leaderboard: A Competitive Benchmark for Safer AI
- Training language models to follow instructions with human feedback
- [2203.02155] Training language models to follow instructions with human feedback
- Carroll Wainwright — Departure from OpenAI (2024) · Ethical AI Departures
- A Right to Warn about Advanced Artificial Intelligence