Google-backed AI network aids lead pipe cleanup in Flint, MI
- To help with the effort in Flint, MI to replace lead pipes that have contaminated residents’ water supplies, researchers from Georgia Tech University developed an artificial intelligence (AI) system, backed by Google, to predict which of the city’s 55,000 homes are affected, according to a report in New Scientist.
- After digitizing old city plans and more than 140,000 handwritten building records, the team analyzed 71 different pieces of information for each property, like the age and location of a home or its values. That has allowed them to target homes that are likely to have lead pipes, with a 97% success rate.
- The system has saved an estimated $10 million by avoiding unnecessary searches, enough to protect an additional 2,000 properties.
The Flint water crisis — which arose after city planners began getting the water supply from the Flint River and the more corrosive water leached lead from old pipes into resident’s supplies — captured headlines and drew attention to aging water infrastructure across the country. A 2016 USA Today investigation published in the aftermath of the Flint crisis found excessive lead levels in 2,000 water systems across every state, collectively covering 6 million people; a 2018 study by the Natural Resources Defense Council also found water systems serving a collective 77 million people had Safe Drinking Water Act violations. Aging pipelines put lead into the water, sometimes with rates higher than the Environmental Protection Agency’s “action level,” where states are required to take action.
Cleaning up those systems has been a priority for both the Obama and Trump administrations; former Trump EPA head Scott Pruitt talked about a "war on lead" that would involve revamping infrastructure to eliminate lead in the water supply in a decade. The EPA is also updating drinking water standards for lead contamination (however, the White House’s budget proposals have cut programs that would clean up water pollution, including lead contamination). But the cost of such a program is astronomical; E&E News quoted public health experts as estimating a wholesale replacement could cost $30 billion.
A major problem is simply finding which homes are served by lead-free pipes, since different standards have been used for decades and most pipes designated as lead-free actually had up to 8% lead until 2014. Michael McDaniel, who initially led Flint’s pipe replacement effort, told New Scientist “we had no good way of doing it,” and said about 20% of the pipes dug up were not made of lead. Other cities would want to avoid wasting money on unnecessary construction and inspection. The AI system can be applied to other cities, its creators said, which should help speed up lead pipe cleanup in other cities — and help give a better idea of the scope of the problem.
Trust issues, however, remain in Flint and other cities, and it will take a substantial campaign to ensure that consumers know their water is safe. Terry Leeds, director of KC Water in Kansas City, MO, last year told Smart Cities Dive that one bright side was that "Flint did allow us to talk about our water more than what we would have had the opportunity to do if it hadn’t happened."
- New Scientist AI hunts down lead pipes in Flint
- Smart Cities Dive 'I still worry': City utility leaders reflect on impact of Flint water crisis
Follow Jason Plautz on Twitter