London Lowmanstone IV

London Lowmanstone IV

Artificial Intelligence Researcher working to improve AI's adaptability to local communities.

London Lowmanstone IV, M.S. - PhD Candidate

My Resume

  • B.A. w/ Honors in Computer Science and Philosophy from Harvard University
  • M.S. in Computer Science from the University of Minnesota
  • PhD Candidate in Computer Science at the University of Minnesota
  • Advised by Maria Gini and Serguei Pakhomov

Hi! I’m a 4th year PhD student at the University of Minnesota focused on building safe artificial intelligence (AI) by building agents that can reason and explain their reasoning about how their actions may be harmful or helpful.

I focus especially on AI systems that need to work in environments where not everyone agrees on what it means to be helpful or hurtful, and the AI needs to be able to reason about multiple different approaches at the same time. This paradigm is what I call pluralistic-values reasoning, which is my primary research focus.

From junior high through college, I worked primarily on AI robotics systems, with an emphasis on natural disaster search and rescue. During college, I realized that the most impactful AI systems would need to be able to communicate with humans, so I came into my PhD program focusing on natural language processing (NLP).

Throughout the course of my PhD, I’ve worked on projects involving logical reasoning (paper) and pluralistic-values prediction models (paper #1, paper #2), and have been a facilitator for many AI safety courses (course #1, course #2). In these last few years of my PhD program, I’m using the expertise I’ve gained to build frameworks and evaluation methods that promote pluralistic-values reasoning for current and future AI systems.