Chemistry

AI analysis shows chemical research often contains inaccurate mass measurement data

Examples of accurate mass calculations and miscalculations using ChemDraw. Credit: Organic Letters (2024). DOI: 10.1021/acs.orglett.4c03458

AI-powered data analysis tools have the potential to significantly improve the quality of scientific publications. A new study by Professor Matthias Christmann, professor of chemistry at the Free University of Berlin, reveals shortcomings in chemical publications.

Christmann analyzed more than 3,000 scientific papers published in Organic Letters over the past two years using Python scripts developed with the latest AI language models. The analysis revealed that only 40% of chemical research papers contain error-free mass measurements. AI-based data analysis tools used for this purpose can be created without any prior programming knowledge.

“These results demonstrate how powerful AI-powered tools can be in everyday research. These tools not only provide access to complex analyses, but also improve the reliability of scientific data. ”explains Christmann.

Advanced large-scale language models such as ChatGPT (OpenAI), Gemini (Google), and Claude (Anthropic) have made it possible to directly translate natural language into computer languages ​​such as Python. This allows researchers with no coding experience to create applications that search for specific text components or measurements in large datasets, for example. The data obtained in this way is further automatically processed and checked for validity.

Christmann’s study, “What we learned from a precise high-volume analysis of 3000 supporting information files,” published in Organic Letters, uses AI-powered data analysis tools to discover previously unknown I found a systematic error. We also identified cases where incorrectly calculated values ​​appeared to be verified by measurements.

“These observations raise the question whether some measurements may have been fabricated,” the researchers stress.

This study shows how AI tools can enhance scientific integrity through automated quality control and systematic error detection.

As part of the “AI in Education” initiative, the Faculty of Biology, Chemistry and Pharmacy at Freie Universität Berlin plans to integrate these and similar tools into the curriculum. “It helps students develop strong data analysis skills and critical thinking abilities,” Christman says. “AI tools will help prepare students for research careers.”

Further information: Mathias Christmann, What I Learned from Analyzing Accurate Mass Data of 3000 Supporting Information Files, Organic Letters (2024). DOI: 10.1021/acs.orglett.4c03458

Provided by Free University of Berlin

Retrieved January 20, 2025 from https://phys.org/news/2025-01-chemical-inaccurate-mass-ai-analysis (January 20, 2025) Studies often include inaccurate mass measurement data. html

This document is subject to copyright. No part may be reproduced without written permission, except in fair dealing for personal study or research purposes. Content is provided for informational purposes only.

Related Articles

Leave a Reply

Your email address will not be published. Required fields are marked *

Back to top button