The Challenge
Facility inspections represent a key enforcement practice across regulatory domains, but are costly, infrequent and detect only a small fraction of violations. Despite increasingly scarce government resources, regulators can improve compliance by adopting data-driven inspection targeting practices and identifying violators more efficiently. Technological advances in computing capabilities, big data availability, and artificial intelligence (AI) can transform the way agencies implement environmental enforcement by together providing reliable predictions of which facilities are most likely to be in violation.
Our Approach
The E&E Lab has pioneered the use of AI to harness the power of data routinely collected by agencies and cutting-edge remote sensing technologies to predict which facilities are most likely to be in violation. This innovative approach complements the expertise of environmental regulators and can deliver significant improvements in environmental protection and human health without increasing regulatory costs.
Since 2015, the E&E Lab has worked with the U.S. Environmental Protection Agency (EPA) to test the ability of machine learning to improve enforcement targeting. Lab researchers, including an embedded E&E Lab fellow at the EPA, developed a robust model to harness administrative data spanning nearly two decades to predict which facilities are most likely to violate hazardous waste regulations. Although the model exceeded the EPA’s historical hit rate, regulators remained unconvinced. So, the E&E Lab proposed a figurative horse race—pitting model-chosen inspections against EPA-chosen ones. The field experiment showed that the model had an 82 percent higher detection rate than the EPA’s status-quo practices, with no increase in cost. Inspired by this success, the EPA decided to roll out the AI targeting model across the hazardous waste program nationwide.
Opportunities for Scale
With E&E Lab’s partnership, the EPA is leading a culture change towards increased reliance on data and AI methods to inform evidence-based decision making. The E&E Lab is harnessing this opportunity to build on the success of the hazardous waste model and apply the machine learning approach to new domains. It is doing so in partnership with the EPA’s newly established National Targeting Center, which offers predictive analytics models to improve targeting as a cornerstone service to environmental regulators at the regional and state levels. Lab researchers are also developing models to target inspections of facilities regulated under the Clean Air Act and the Clean Water Act. Both models leverage novel remote sensing techniques to detect when and where pollutant emissions occur, representing a pioneering way to use satellite data to support inspection targeting. And, the Lab has partnered with state environmental agencies to improve targeting for a variety of environmental challenges—from methane emissions to ozone pollution. The E&E Lab’s AI models are helping environmental agencies across the country address their unique priorities through improved inspection targeting.
