Critical water-quality tests, such as measuring BOD, take up to five days to process in a lab. However, wastewater treatment plant operators routinely need to make chemical dosing and treatment decisions within hours to protect aquatic life and meet regulatory standards.
While artificial intelligence has the computing potential to predict these levels instantly using historical data, operators are sometimes hesitant to use it because they can't see the process for how it arrives at its prediction. In other words, it lacks the transparency required for high-stakes municipal water management.
But research by Fuad Nasir, a water supply specialist with the Wisconsin Department of Natural Resources and a University of Wisconsin-Milwaukee graduate, highlights a practical breakthrough called Explainable AI (XAI). This technology aims to bridge the trust gap by predicting outcomes while simultaneously displaying the exact variables, such as temperature, flow rate or ammonia levels, that are driving those predictions. Nasir used XAI in his dissertation at UWM, and now he uses it in his work at the DNR.
By allowing operators to visualize the data, XAI enables immediate data-driven operational responses. While the technology is gaining traction in the sector, bringing XAI to the municipal level will still require utility investments in modernized instrumentation, software training and targeted funding.