Aligned with Whom? Direct and social goals for AI systems
It addresses the AI alignment problem by clarifying goal types for AI developers and policymakers, but is incremental as it builds on existing alignment discussions.
The paper distinguishes between direct alignment, where AI systems pursue the operator's goals, and social alignment, which considers broader societal impacts and externalities, proposing that solutions involve robust implementation for direct alignment and governance and norm design for social alignment.
As artificial intelligence (AI) becomes more powerful and widespread, the AI alignment problem - how to ensure that AI systems pursue the goals that we want them to pursue - has garnered growing attention. This article distinguishes two types of alignment problems depending on whose goals we consider, and analyzes the different solutions necessitated by each. The direct alignment problem considers whether an AI system accomplishes the goals of the entity operating it. In contrast, the social alignment problem considers the effects of an AI system on larger groups or on society more broadly. In particular, it also considers whether the system imposes externalities on others. Whereas solutions to the direct alignment problem center around more robust implementation, social alignment problems typically arise because of conflicts between individual and group-level goals, elevating the importance of AI governance to mediate such conflicts. Addressing the social alignment problem requires both enforcing existing norms on their developers and operators and designing new norms that apply directly to AI systems.