My biggest contribution is a definition of "human values", “meaningful living”, etc that's precise enough to create surveys, metrics, aligned ML models, etc. That same formalization of human values can enable explainable, moral learning in AI, and be used to design systems that support meaningful living, perhaps addressing problems with mechanisms like markets, ML, or social media, which optimize for engagement and revealed preference, rather than values.

This work is a formalization of definitions of values from Ruth Chang, Charles Taylor, and others.