Tech Explained: California’s Water Agencies Begin Looking to AI in Simple Terms

This is the second in a multi-part series examining how artificial intelligence may impact California water.

Water is known for being a cautious sector. While pressure grows on some workers to use AI for more tasks, most California water agencies are just beginning to take advantage of the technology. Eventually, AI is likely to help water agencies with a range of applications, including finding ways to save time, reduce water use, and bring down costs.

So how are California water agencies currently using AI, and what should agencies consider as they adopt the technology? We spoke with experts who shared some key first steps.

1. Have conversations now to identify how AI can help your agency
As AI applications become more ubiquitous, experts caution that organizations should be thoughtful about where AI can add value. Chris Tull, chief data officer of the California Data Collaborative, and Jim Cooper, global director of Water Optimization at Arcadis, agree that this is an important moment for water agencies to figure out how AI can be useful. As Cooper said, “AI can access untapped insights and benefits. However, you can also ask ChatGPT to add 1+1 for you, but should you use it for that?”

In our research, we found that agencies are currently using AI for a range of purposes:

• Optimizing wastewater treatment. The Eastern Municipal Water District (EMWD) in Riverside County has applied machine learning models to reduce the energy demands of wastewater treatment, which is expected to save the agency an estimated $100,000 in costs each year.
• Enabling smart pipeline maintenance. EMWD is using AI to detect leaks and to identify and prioritize pipes for repair or replacement, which can save costs by reducing water lost before it reaches a customer. In Arizona, Tucson Water has partnered with VODA.ai to predict pipeline failures.
• Modeling water use. The Moulton Niguel Water District in Orange County is using machine learning with advanced metering infrastructure to better model water use and detect failing meters.
• Delivering real-time alerts and awareness. The Los Angeles Department of Water and Power has implemented a platform that uses AI to integrate diverse operational data into centralized dashboards. These dashboards give staff real-time information and help inform management decisions.

2. Ensure that high-quality data are being collected
Agencies report that it’s important to have high-quality data that is “AI-ready.” For instance, EMWD began tracking leaks in 2005, which gave it a body of data that allowed it to evaluate new machine learning models, said the agency’s director of maintenance, Dave Brown, in a podcast. Moulton Niguel is currently investing in preparing and integrating data from different sources and ensuring its quality.

3. Set organizational policies around appropriate use, including protecting customer privacy
Water agencies should get policies and guidelines in place to mitigate possible risks from AI, from protecting customer privacy to mitigating cyber security risks. The GovAI coalition, created by the city of San Jose, focuses on promoting responsible AI in the public sector. It provides water districts with template AI policies to help them get started.

4. Train workers to use AI effectively and appropriately
While a case study of DC Water reports large productivity and efficiency gains with Microsoft’s Copilot AI tool, the study also found that staff training was essential. In California, a coalition of Orange County agencies recently launched the first AI course targeted to water professionals.

What’s next?

What about the risk that’s top of mind for many: will AI take my job? Based on our conversations, AI doesn’t seem to be taking over jobs in the water sector—yet. This is in part because, at water agencies, there is often more work than bandwidth; early applications of AI are expected to increase productivity without replacing people. EMWD’s Dave Brown explained that while AI is helping them identify leaks, they still need human workers to go out to address those leaks.

There are signs that greater application of AI in water agencies will be coming soon. The Metropolitan Water District released a request for information last fall looking for companies to evaluate the capabilities and accuracy of different AI products.

Water agencies may be feeling pressure to adopt AI quickly. New coalitions and training courses can help them prepare. As the sector embraces this technology, thoughtful application, clear policies, ongoing data collection, and training will be key.