Home / Current Issue / Paper 1722830
Beyond Water Quality Monitoring: Toward AI-Enabled Early Detection of Hidden Environmental Health Risks from Emerging Contaminants
Subject area: Science,Engineering and Technology · Area of research: Technology
Abstract
Water quality monitoring relies on a fixed, regulated list of things to test for, measured against limits set decades ago. But the chemicals actually present in water and in the human body are far more numerous, and increasingly dominated by substances routine testing does not cover, including PFAS ("forever chemicals"), pharmaceutical residues, microplastics, endocrine-disrupting chemicals, and pesticide breakdown products. This review brings together research from analytical chemistry, exposure science, and artificial intelligence (AI) to ask a question no existing review has directly answered: can AI move environmental health monitoring beyond routine compliance testing toward genuinely detecting hidden risks, meaning contaminants that are chemically real and biologically plausible but currently invisible to regulators? We examine four areas where AI is being used, finding unknown chemicals, real-time sensor monitoring, linking exposure to biological effects, and forecasting contamination, and identify where the field agrees and where claims remain unproven. The evidence is mostly early-stage: small datasets, testing on the same data used to build the model, limited interpretability, and little connection to real health outcomes. Building on these gaps, we propose a three-step AI-Enabled Sentinel Framework that keeps chemical detection separate from health-risk assessment and requires models to be interpretable and independently tested from the outset. We close with a research agenda and implications for regulators, utilities, and communities facing unequal exposure. Our central point: the next stage of water quality monitoring will come not from testing the same known list more precisely, but from AI systems that can responsibly flag the unknown, once the field closes the gaps described here.
Keywords
emerging contaminants; artificial intelligence; machine learning; early warning systems; exposome; non-target screening; environmental health surveillance; water quality
How to cite this paper
@article{1722830,
author = {Ishola Abdul Dimeji, Jumoke Oluwafunmito Omotosho, Linda Egbubine, Adejoke Olowookere-Oduyebo},
title = {Beyond Water Quality Monitoring: Toward AI-Enabled Early Detection of Hidden Environmental Health Risks from Emerging Contaminants},
journal = {Iconic Research And Engineering Journals},
year = {2026},
volume = {10},
number = {3},
pages = {681-691},
issn = {2456-8880},
url = {https://www.irejournals.com/formatedpaper/1722830.pdf},
abstract = {Water quality monitoring relies on a fixed, regulated list of things to test for, measured against limits set decades ago. But the chemicals actually present in water and in the human body are far more numerous, and increasingly dominated by substances routine testing does not cover, including PFAS ("forever chemicals"), pharmaceutical residues, microplastics, endocrine-disrupting chemicals, and pesticide breakdown products. This review brings together research from analytical chemistry, exposure science, and artificial intelligence (AI) to ask a question no existing review has directly answered: can AI move environmental health monitoring beyond routine compliance testing toward genuinely detecting hidden risks, meaning contaminants that are chemically real and biologically plausible but currently invisible to regulators? We examine four areas where AI is being used, finding unknown chemicals, real-time sensor monitoring, linking exposure to biological effects, and forecasting contamination, and identify where the field agrees and where claims remain unproven. The evidence is mostly early-stage: small datasets, testing on the same data used to build the model, limited interpretability, and little connection to real health outcomes. Building on these gaps, we propose a three-step AI-Enabled Sentinel Framework that keeps chemical detection separate from health-risk assessment and requires models to be interpretable and independently tested from the outset. We close with a research agenda and implications for regulators, utilities, and communities facing unequal exposure. Our central point: the next stage of water quality monitoring will come not from testing the same known list more precisely, but from AI systems that can responsibly flag the unknown, once the field closes the gaps described here.},
keywords = {emerging contaminants; artificial intelligence; machine learning; early warning systems; exposome; non-target screening; environmental health surveillance; water quality},
month = {September},
}