An AI-Assisted Review of Trends in Environmental and Human Health

Summary

This narrative review uses generative artificial intelligence (AI) tools based on large language models (LLMs) to summarize scientific literature related to environmental and human health. It evaluates three distinct methodologies for knowledge synthesis using LLMs and investigates applications of informatics in citizen science, environmental health, and drug discovery.

Three approaches were tested. The first approach was a reading list-based review using preselected citations and Google NotebookLM to assess machine learning and citizen science applications to environmental health. The second approach was an AI-assisted narrative review utilizing scite.ai to synthesize data on drug discovery from natural environmental sources. The third approach was a conversational LLM-based synthesis employing ChatGPT to explore bioinformatics in marine and medical sciences, generating new research questions and thematic areas.

AI tools helped summarize the literature and conclude that machine learning combined with citizen science improves environmental data collection and analysis through the use of low-cost sensors and public data annotation efforts. On the drug discovery end, natural products remain a vital source, with advancements in genome mining and cheminformatics enhancing research efficiency. Finally, marine bioinformatics links environmental and human health, identifying shared challenges like pollution impacts and microbial resilience.

Generative AI tools demonstrate potential in scientific literature synthesis as evidenced by their ability to identify relevant application of data science and informatics in environment and human health applications. Some challenges, such as some omitted references and the need to iterate through the prompts necessitate the need for expert supervision of these tools.