Leveraging generative AI for causal inference with unstructured data
- 1Department of Government, Harvard University
- 2Department of Statistics, Harvard University
- 3John F. Kennedy School of Government, Harvard University
Abstract
We introduce GenAI-Powered Inference (GPI), a statistical framework for causal inference using unstructured data, including text and images. GPI leverages open-source pretrained Generative AI (GenAI) models-such as large language models and diffusion models-not only to generate unstructured data at scale but also to extract low-dimensional representations that are guaranteed to capture their underlying structure. Applying machine learning to these representations, GPI enables estimation of causal effects while quantifying estimation uncertainty. Unlike existing approaches to representation learning, GPI does not require fine-tuning of GenAI models, making it computationally efficient and broadly accessible. We illustrate the versatility of the GPI framework through three applications: 1) estimating the effects of Chinese social media censorship while adjusting for textual confounders, 2) isolating the impact of specific image features from that of other correlated features in the same image, and 3) assessing the persuasiveness of political rhetoric. An open-source software package is available for implementing GPI.
Related
- GenAI Powered Dynamic Causal Inference with Unstructured Data
- Causal Inference with Generative Artificial Intelligence: Application to Texts as Treatments
- Generative Marketing Mix Modeling: A Causal Inference Framework Linking GEO and GEM to Business Impact
- Identifying ODEs from Unstructured Data with Causal Representation Learning
- Semiparametric inference for a Target Outcome Leveraging Heterogeneous External Summaries of Auxiliary Outcomes
BibTeX
@article{imai2026leveraging,
title = {{Leveraging generative AI for causal inference with unstructured data}},
author = {Kosuke Imai and Kentaro Nakamura},
journal = {Proceedings of the National Academy of Sciences},
year = {2026},
month = sep,
doi = {10.1073/pnas.2530532123},
url = {https://doi.org/10.1073/pnas.2530532123},
note = {CC BY-NC-ND},
}
RIS
TY - JOUR
TI - Leveraging generative AI for causal inference with unstructured data
AU - Kosuke Imai
AU - Kentaro Nakamura
PY - 2026
DA - 2026/09/03
JO - Proceedings of the National Academy of Sciences
DO - 10.1073/pnas.2530532123
UR - https://doi.org/10.1073/pnas.2530532123
AB - We introduce GenAI-Powered Inference (GPI), a statistical framework for causal inference using unstructured data, including text and images. GPI leverages open-source pretrained Generative AI (GenAI) models-such as large language models and diffusion models-not only to generate unstructured data at scale but also to extract low-dimensional representations that are guaranteed to capture their underlying structure. Applying machine learning to these representations, GPI enables estimation of causal effects while quantifying estimation uncertainty. Unlike existing approaches to representation learning, GPI does not require fine-tuning of GenAI models, making it computationally efficient and broadly accessible. We illustrate the versatility of the GPI framework through three applications: 1) estimating the effects of Chinese social media censorship while adjusting for textual confounders, 2) isolating the impact of specific image features from that of other correlated features in the same image, and 3) assessing the persuasiveness of political rhetoric. An open-source software package is available for implementing GPI.
SN - 0027-8424
SN - 1091-6490
ER -