My research centers on the value alignment process in artificial intelligence, focusing on building systems that appropriately adhere to our social and moral values in both their operation and impacts. I am interested in exploring the social and technical dimensions of this challenge.

In the context of my research, our values determine our preferred state of the world and the relation between these preferences across different contexts. Technology can be argued to have a reflexive relationship with these values, both shaping and being shaped by them. Artificial intelligence is no exception, but its autonomous and opaque nature poses new challenges in understanding and controlling how our values influence its behavior and, in turn, how it affects us.

From a social perspective, my goal is to understand how the AI design process impacts human values and the effects of AI systems on the values of the humans interacting with them. I aim to use this understanding to improve the AI design process, ensuring it better reflects all stakeholder values, not just those of the developers, and to control for undesirable value impacts from the adoption of this technology.

On the technical side, I am interested in methods to represent and reason over values. Given that values are abstract concepts with complex interrelations, formalizing them in a way that is interpretable by AI systems is a significant challenge. I am exploring both mathematical structures and logic-based representations that can encode values in a way that is useful for both human users and AI systems.

My other research interests include augmented cognition (particularly for supporting cooperation and understanding), explainable AI, and AI governance.