About

Hello, I'm Gabe Mukobi! I am a technical staff member at the U.S. AI Safety Institute developing AI safety infrastructure and standards and an incoming Ph.D. student in Computer Science at UC Berkeley advised by Jacob Steinhardt and Dawn Song. My research focuses on technical AI governance, especially understanding, securing, and preventing the misuse of language model systems. Read my list of current research projects here and my agenda of research questions here.

Previously, I was a Technology and Security Policy fellow at RAND working on U.S. AI national security policy. I completed my M.S. and B.S. in Computer Science at Stanford, where I founded Stanford AI Alignment and SPAR.

My other interests include animal welfare, games, music, film, 3D art, photography, virtual reality, fantasy, and tea!

Gabriel Mukobi

Highlighted Publications:

Escalation Risks from Language Models in Military and Diplomatic Decision-Making

Juan-Pablo Rivera*, Gabriel Mukobi*, Ann-Katrin Reuel*, Max Lamparth, Chandler Smith, Jacquelyn Schneider

Accepted to the ACM FAccT 2024, MASEC NeurIPS 2023 workshop.

Welfare Diplomacy: Benchmarking Language Model Cooperation

Gabriel Mukobi*, Hannah Erlebach, Niklas Lauffer, Lewis Hammond, Alan Chan, Jesse Clifton

Accepted to the SoLaR NeurIPS 2023 workshop.

Societal Adaptation to Advanced AI

Jamie Bernardi*, Gabriel Mukobi*, Hilary Greaves*, Lennart Heim, Markus Anderljung*.

SuperHF: Supervised Iterative Learning from Human Feedback

Gabriel Mukobi*, Peter Chatain*, Su Fong*, Robert Windesheim, Gitta Kutyniok, Kush Bhatia, Silas Alberti

Accepted to the SoLaR NeurIPS 2023 workshop.

The WMDP Benchmark: Measuring and Reducing Malicious Use With Unlearning

Nathaniel Li*, Alexander Pan*, ..., Gabriel Mukobi, ..., many co-authors

Accepted to ICML 2024.

Open Problems in Technical AI Governance

Anka Reuel*, Ben Bucknall*, ..., Gabriel Mukobi, ..., many co-authors

In-Progress Publications:

AI Consciousness and Public Perception: Four Futures

TBA co-authors, Gabriel Mukobi

TBA Science of Evaluations Paper

Gabriel Mukobi, TBA co-authors.

Portfolio Sites

Sticks and Stones Software website screenshot

Software Portfolio

Come to my software development portfolio to see some of the AI projects, games, websites, and desktop software I create!

Digital 3D Art

My ArtStation portfolio is where I post 3D art made with Blender, Houdini, Unreal Engine, and more!

ArtStation Porfolio Thumbnail
photography portfolio screenshot

Photography

A minimalistic photography portfolio arranged in a neat dynamic grid layout.

Mukobi Music

A website for anything regarding the musical side of me, including information about my music, groups I've played with, and more.

Mukobi Music website screenshot
film portfolio screenshot

Film

A minimalistic film portfolio to highlight short films, video essays, and other motion picture works I've made.

Contact

Please use one of the following methods to contact me. I will try to get back to you as soon as possible!