Towards a Planetary Health Impact Assessment Framework: Exploring Expert Knowledge & Artificial Intelligence for RF-EMF Exposure Case-Study

Authors: Stefanopoulou M, Sonnenschein TS, de Gannes FP, Scheider S, Vermeulen R, Röösli M, Huss A

Year: 2025 Dec

Category: Environmental Health, Electromagnetic Field Safety

Journal: Bioelectromagnetics

DOI: 10.1002/bem.70038

URL: https://onlinelibrary.wiley.com/doi/10.1002/bem.70038

Abstract

Overview

While recent World Health Organization (WHO) systematic reviews have extensively assessed the direct health effects of radiofrequency electromagnetic field (RF-EMF) exposure, the possible indirect impacts on human health through ecosystem disruption have remained unstudied. This gap prompted the proposal of a Planetary Health Impact Assessment (PHIA) approach, which integrates both direct and ecologically mediated exposure pathways.

Methods

This study explores the construction of a PHIA framework using knowledge graphs (KGs) to organize and visualize complex, interdisciplinary knowledge. The case study focused on RF-EMF exposure from mobile telecommunication technologies, utilizing both expert input and artificial intelligence (AI) tools that incorporate Natural Language Processing (NLP) and Deep Learning.

  • Expert-based KGs were developed collaboratively with 12 specialists to hypothesize pathways linking RF-EMF exposure to direct and indirect health effects.
  • The AI-based tool rapidly extracted and visualized information from scientific literature into KGs, though expert validation was necessary due to AI's limitations in precision and context.

Findings

  • Experts jointly visualized pathways from RF-EMF exposure to direct health effects on organisms and indirect effects on humans via ecosystem disruption.
  • AI-assisted KGs identified varying structures but require substantial human oversight for accuracy.
  • The study found gaps in the literature, especially regarding ecological effects on pollinators, birds, and plants and subsequent impacts on human health.
  • Data compiled: 97 publications on RF-EMF's potential effects on organisms/humans and 13 reviews on ecological consequences, resulting in 4215 unique direct association instances and 232 indirect.
  • Visualization highlighted links between RF-EMF exposure, species diversity, community health of pollinators, and broader ecosystem services, connecting further to human health implications.

Conclusion

The expert-based knowledge graph serves as an organizer of available knowledge and a preliminary tool for PHIA development. AI tools show promise for enhancing exploratory analysis of literature, but human judgment remains irreplaceable for context-sensitive validation. The study underscores the potential for both direct and indirect health risks from electromagnetic fields, highlighting the need for further research into ecological and human health linkages.

RF-EMF exposure planetary health knowledge graph AI tool ecological impact indirect health risk environmental assessment
← Back to Stats