In-Context Credit Assignment via the Core
This work addresses the problem of fair credit assignment for AI-generated content, which is important for creators and platforms, but the results are incremental as they apply existing game theory concepts with algorithmic improvements.
The paper introduces a cooperative game theory approach for assigning credit to content creators whose work appears in AI-generated outputs, using the least core solution concept. Their algorithms approximate the least core with orders of magnitude fewer LLM calls than alternatives on a web retrieval task.
We propose incentive-aligned mechanisms for in-context credit assignment: the task of assigning credit for AI-generated content (e.g. code, news articles, short-form videos) among creators whose intellectual property appears in the context window. Our approach is based on the least core solution concept from cooperative game theory, which distributes value in a way that is as stable as possible by ensuring that no subset of creators is significantly under-compensated relative to the value they could generate on their own. We develop algorithms for approximating the least core, which leverage novel routines for constraint seeding and constraint separation. On a web retrieval credit assignment task, we find that our approaches are capable of approximating the least core using orders of magnitude fewer LLM calls compared to alternative methods.