A Multi-Level Framework for the AI Alignment Problem
This work addresses the normative challenge of encoding human values in AI systems, but it is incremental as it provides a conceptual framework without new empirical results.
The paper tackles the AI alignment problem by proposing a multi-level framework that examines value alignment at individual, organizational, national, and global levels, illustrating how values interact across these levels and applying it to AI content moderation.
AI alignment considers how we can encode AI systems in a way that is compatible with human values. The normative side of this problem asks what moral values or principles, if any, we should encode in AI. To this end, we present a framework to consider the question at four levels: Individual, Organizational, National, and Global. We aim to illustrate how AI alignment is made up of value alignment problems at each of these levels, where values at each level affect the others and effects can flow in either direction. We outline key questions and considerations of each level and demonstrate an application of this framework to the topic of AI content moderation.