Can LLM Automate Causal Inference? | Causal Summit 2026 | Pr. Nick Huntington-Klein

Nick Huntington-Klein — Can LLMs Automate Causal Inference? Evidence from the Coronary Drug Project (Block 5 · Causal AI & LLMs) Can LLMs automate the process of causal inference by providing the information about causal links necessary to build a structural model? Using the case of confounding in the Coronary Drug Project (CDP) — for which several studies list expert-selected confounders that can serve as a ground truth — Nick shows that LLMs exhibit mediocre performance in identifying confounders in this setting, even though text about the ground truth is in their training data. Further, LLM judgment on confounder status is highly inconsistent across models, prompts, and irrelevant concerns. LLMs do not yet have the ability to automate the reporting of causal links. About the speaker: Nick Huntington-Klein is an associate professor of economics at Seattle University. He is well-known for his work on metascience, education policy, and — through his textbook The Effect — causal inference. His work has been published in Economic Inquiry, the Journal of Economic Behavior and Organization, and the Journal of Causal Inference. Links: The Effect (textbook): https://theeffectbook.net/ LinkedIn:   / acoaaalog2ibgzycgc84eqj8i2-srhnpxdhgwbk   More about The Causal Summit: https://causalsummit.com