Sarah H. Cen
I am an Assistant Professor at Carnegie Mellon University in the Departments of Electrical & Computer Engineering and Engineering & Public Policy. My research falls at the intersection of artificial intelligence, society, and law. On the technical side, I love learning new methods. I use tools from a wide range of fields, including machine learning, statistics, game theory, causal inference, and information theory to tackle research questions. I believe in picking the right research questions first, then using the right tools to answer them. The goal of my research is to better design, understand, and govern AI and automated systems. As a consequence, I also have a strong interest in law, policy, ethics, and philosophy. Recently, I have written on taxing AI, the AI supply chain, barriers to evidence in AI-related litigation, AI auditing, AI & elections, and social media regulation.
I recently finished my postdoc at Stanford HAI, where I had an amazing time working with Percy Liang in Computer Science and Daniel E. Ho in the Law School's RegLab. Before that, I completed my PhD at MIT EECS under the wonderful mentorship of Aleksander Mądry and Devavrat Shah, my master's at Oxford with Paul Newman, and my undergraduate at Princeton with Naomi Leonard.
A few recent highlights:
- Check out some recent work mapping the AI Supply Chain on this website here (additional resources: working draft on mapping the AISC, early paper on the AISC, a more recent paper on AI entanglement, an earlier blog series)
- New paper on Taxing Artificial Intelligence here. We did not start with the assumption that taxing AI would be a good idea. We researched taxation, the options, the upsides and downsides, then put it all in one place.
- Ran a large-scale longitudinal audit of language models during the 2024 US Election here.
Please find my (always, slightly out-of-date) C.V. here.