Toward a Causal Data Management Ecosystem for Decision Making and Agentic AI

Modern AI is an ecosystem of ML predictors, multimodal models, LLMs and agents whose outputs feed each other across heterogeneous, independently governed data sources. Integration alone is not enough: correlational signals invite confounded decisions, and autonomous agents must anticipate the consequences of their actions. We argue the integrated ecosystem needs an explicit causal layer, built as a shared, persistent, queryable Causal World System (CWS).

arXiv: 2608.07214

Recommended citation: D. Qiu, Y. Zhou, A. Pachera, A. Bonifati, A. Mauri. Toward a Causal Data Management Ecosystem for Decision Making and Agentic AI. ACM AI Leadership Summit 2026, Atlanta. arXiv:2608.07214
Download Paper