# Evan Hubinger

Research and product · Alignment Science org lead at Anthropic · joined 2023

Co-wrote the 2019 paper on risks from learned optimization the year he graduated. He now leads Anthropic's alignment science group.

Full profile: https://www.autumn.ai/resources/anthropic/people/evan-hubinger

**At Anthropic:**
- 2026 – present: Leads the Alignment Science org
- 2024–2025: Led the Alignment Stress-Testing team
- 2023: Alignment Science researcher
Sources: [1](https://www.linkedin.com/in/ehubinger)

**Career before Anthropic:**
- Nov 2019 – Jan 2023: Machine Intelligence Research Institute, Research Fellow
Sources: [1](https://www.linkedin.com/in/ehubinger)

**Education:**
- Bachelor's Degree, Mathematics and Computer Science, Harvey Mudd College, 2019
Sources: [1](https://www.linkedin.com/in/ehubinger)

**Lesser-known facts:**
- Co-wrote the Sleeper Agents and Alignment Faking papers. [1](https://www.alignmentforum.org/users/evhub)
- Created the Coconut programming language. [1](https://www.linkedin.com/in/ehubinger)
- Interned on OpenAI's safety team in 2019 under Paul Christiano, working on amplification and universality. [1](https://evhub.github.io/resumes/ehubinger_resume_main.pdf)
- Built a deep reinforcement learning agent for HRL Laboratories to tune quantum dots while at Harvey Mudd. [1](https://evhub.github.io/resumes/ehubinger_resume_main.pdf)
- Interned as a software engineer at Ripple and Yelp and in site reliability at Google before turning to alignment. [1](https://www.linkedin.com/in/ehubinger)
- Wrote "An overview of 11 proposals for building safe advanced AI," comparing each on inner alignment, outer alignment and competitiveness. [1](https://evhub.github.io/resumes/ehubinger_resume_main.pdf)
- Posted in September 2026 that his personal estimate of AI killing all humans within the next decade is above 10%. [1](https://x.com/EvanHub/status/2097497037956891126)
- Wrote in the same thread that Anthropic does "not yet have a plan to solve alignment for superintelligence" and is not clearly on track. [1](https://x.com/EvanHub/status/2097497037956891126)

**Interviews and talks:**
- [FLI AI Alignment Podcast on inner alignment, outer alignment and proposals for safe AI (2020)](https://futureoflife.org/podcast/evan-hubinger-on-inner-alignment-outer-alignment-and-proposals-for-building-safe-advanced-ai/)
- ["How An AI Model Learned To Be Bad" with Monte MacDiarmid, on reward hacking (2025)](https://www.youtube.com/watch?v=lvRxmAV49yI)

**Known for:**
- Risks from Learned Optimization in Advanced Machine Learning Systems (2019)
- Alignment Faking in Large Language Models (2024)
- Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training (2024)
- Model Organisms of Misalignment: The Case for a New Pillar of Alignment Research (2023)
- Alignment remains a hard, unsolved problem (2025)
- Natural emergent misalignment from reward hacking in production RL (2025)
Sources: [1](https://www.alignmentforum.org/users/evhub)

Researched with Autumn (https://www.autumn.ai) from public sources. Updated 2026-10-18. Everyone at Anthropic: https://www.autumn.ai/resources/anthropic/people