As environmental epidemiologists, one of our primary goals is to conduct research where the findings can inform policy and regulatory action to promote public health. To effectively do so, our findings need to be robust and not impacted by different sources of bias. Recently, there has been increased interest in the use of ‘causal inference modeling’, a class of models that aim to approximate randomized experiments and are more robust to model misspecification compared to more traditional approaches. Although causal inference modeling is not new, recent methodological advances allow the use of continuous and high-dimensional exposures, allowing environmental epidemiologists to begin more broadly applying these models. However, with increasing applicability of causal inference modeling, the misconception that these models (and only these models) can infer causality also arose. In an extreme realization of this misconception, the US Environmental Protection Agency’s Clean Air Scientific Advisory Committee (notably without any epidemiologist members) recently proposed that only papers using such methods should be used to inform pollutant standards and all previous papers that did not should be discarded. This proposal was met with strong opposition from the scientific community. In this ISEE North American Chapter-sponsored symposium, we will discuss five different applications that use causal inference modeling to evaluate associations between different exposures (air pollution, arsenic, fluoridation, antibiotic use in livestock feed, and coal-fired power plant-emitted metals) and adverse health outcomes aiming to inform policy and/or evaluate existing or proposed policies. Importantly, we will discuss the advantages and disadvantages of causal inference modeling for environmental policy, what we can and cannot infer given the results of using such models, and how we can best strengthen evidence for causality.