We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
Leveraging AI tools for microplastic data quality assessment
Summary
This study explored how AI tools can be used to assess data quality in microplastic research, addressing the challenge that studies are conducted at widely varying quality levels. AI-assisted quality assessment identified key sources of uncertainty and bias, providing a scalable approach to evaluate the reliability of the growing body of microplastic exposure and effects literature.
There is an urgent need to assess the risk of microplastics on human health. Increasing research efforts have focused on the exposure and effects of microplastics on the human body. Despite the rapid growth in research progress, studies are conducted at varying quality levels, resulting in uncertainty about the reliability and applicability of their data for risk assessment. Quality assurance and quality control (QA/QC) screening are critical for the human risk assessment process. However, the large volume of publications makes this process laborious. Novel technologies such as natural language processing (NLP) and artificial intelligence (AI) have the potential to address this issue at practice. For instance, OpenAI's ChatGPT chatbot can process large amounts of text data and generate coherent outputs using NLP technology. Its potential in research has been widely discussed, and it has recently been proven useful in text annotation tasks and generating information for medical queries. In this research, we tested the potential to use ChatGPT to conduct the QA/QC screening of existing microplastic research. We prompted ChatGPT with instruction text written based on our previous QA/QC criteria on drinking water. We fed a set of recent published papers to ChatGPT and collected the response. We further compared them with the assessment results from human professionals and computed the accuracy of ChatGPT. Our results demonstrated that generative language AI tools like ChatGPT can be useful in extracting key information and significantly accelerates the QA/QC process. Despite the promising outcomes of ChatGPT in QA/QC screening, we acknowledge that caution is necessary in the early stages of utilizing such innovative technologies, as they may have certain limitations. Adopting a human-in-the-loop approach is essential, for instance by randomly checking subsets of the AI results. Further research is needed to improve the performance and applicability of such AI tools in microplastic research. Also see: https://micro2024.sciencesconf.org/559266/document
Sign in to start a discussion.