Fair Document Valuation in LLM Summaries via Shapley Values

Researchers proposed Cluster Shapley, an approximation method for fairly valuing and compensating content creators whose work is summarized by Large Language Models. The approach groups semantically similar documents via LLM embeddings and computes Shapley values at the cluster level, with formal bounds on approximation and revenue-attribution error. On Amazon product review data, Cluster Shapley "substantially improves the efficiency--accuracy frontier" over Monte Carlo sampling, Kernel SHAP, and simple attribution heuristics. The method is agnostic to the specific LLM and summarization process.