A Board Level Guide to Risks and Opportunities of AI in the Water Sector

A Board Level Guide to Risks and Opportunities of AI in the Water SectorAbstractEarlier this year, we presented a talk to the Water Advisory Committee of Orange County, CA (WACO) about risks, barriers, and opportunities of artificial intelligence in the water sector at a high-level but using specific examples. We set out to answer the question that many board members and executives are asking including 'What is this artificial intelligence stuff and how does it apply to our utility?' and 'What do I need to know about it in terms of risks and use cases?' The talk was designed to be highly accessible to people of all ages and backgrounds and was well-received. With the fast-changing landscape of AI that includes a broad range of applications such as Narrow AI, Generative Design, Process Support, Natural Language Processing, Virtual Assistants and Large Language Models (e.g., ChatGPT) and Robotics, etc., utility managers and board members are looking for guidance on what types of questions to ask and what types of risks they may need to consider. We have observed many cases of people across the water sector using ChatGPT to develop RFPs, to create memos, to better understand technical topics, and to find advice on solving problems specific to the water sector, sometimes leading to erroneous or incomplete advice. In this talk, we spend some time defining artificial intelligence (AI) and machine learning (ML) in the context of applications that we are seeing implemented in water and wastewater utilities. AI and ML applications span various executive, business and operations functions and target various outcomes within these functions. These include technologies that span from ChatGPT to Microsoft CoPilot to natural language processing models. We walk through specific examples ranging from applications supporting business functions such as those targeted at long range CIP planning, to those that are targeted toward achieving everyday operational efficiency of specific unit processes such as chemical feed dosing optimization. The talk explores the level of development, adoption, and validation of the various AI tools. It also delves into risks of relying on the AI design models and potential mitigation of these risks. Machine learning, which is a subset of AI, focuses on the development of algorithms that enable computers to learn and make predictions without explicit programing. This talk also explores opportunities and risks within machine learning and some prominent applications that are being used or are currently being developed for the water industry especially where ML has decisive benefits over traditional modeling techniques. We discuss the various types of ML models: supervised learning, unsupervised learning and reinforcement learning models and the examples of their use cases within the water industry. We examine the role and need of subject matter experts in development of ML models and the potential that can be realized through these applications. Lastly, the talk will provide context of how to set the building blocks within an organizations' framework for a data driven future so that they can begin to leverage AI and ML tools with high quality data sets.This paper was presented at the WEF/AWWA Utility Management Conference, February 13-16, 2024.SpeakerAhuja, NanditaPresentation time09:00:0009:30:00Session time08:30:0010:00:00SessionManagement and Oversight of New Frontiers in Artificial IntelligenceSession number16Session locationOregon Convention Center, Portland, OregonTopicDigital Transformation including AI and ChatGPTTopicDigital Transformation including AI and ChatGPTAuthor(s)Ahuja, NanditaAuthor(s)N. Ahuja1, B. Stanford1, K. Bilyk, J. Roostaei1Author affiliation(s)Hazen and Sawyer 1;SourceProceedings of the Water Environment FederationDocument typeConference PaperPublisherWater Environment FederationPrint publication date Feb 2024DOI10.2175/193864718825159314Volume / Issue Content sourceUtility Management ConferenceWord count15