Beneath the streets of American cities, lead pipes carry both water and an invisible burden of risk—one that falls hardest on communities least equipped to find it. Researchers have now trained an artificial intelligence to read the hidden geometry of water infrastructure, learning from one city's crisis to illuminate risk in another. By treating water systems as what they truly are—networks of shared fate rather than isolated properties—the model achieves meaningful accuracy even where data is scarce. It is a quiet but consequential step toward making an old and stubborn danger visible before
AI model identifies lead contamination risk across cities with limited data
Lead contamination in residential water poses persistent public health risks, particularly affecting vulnerable populations in data-limited jurisdictions.