# ChatGPT Struggles With Science Accuracy, Study Warns
Research: ChatGPT Often Misunderstands Science
Mirage News2026.03.18Updated 27d ago
Washington State University professor Mesut Cicek and his research team repeatedly tested ChatGPT by giving it hypotheses taken from scientific
AI SummaryPowered by AI
Researchers at Washington State University tested ChatGPT's ability to evaluate scientific hypotheses and found the AI frequently misinterprets findings. Professor Mesut Cicek's team discovered significant gaps in the chatbot's scientific comprehension. The findings raise concerns about relying on ChatGPT for research or educational purposes.
Original Article
Read full article on Mirage Newschatgptwashington state universitymesut cicekai accuracyscientific researchmachine learninghypothesis testingresearchoftenmisunderstands
Explore More
Related News
![[Science Bulletin] Korean Research Councils Discuss Bridging Science, Humanities, and National Policy](https://image.dongascience.com/Photo/2026/04/17756412415928.jpg "[Science Bulletin] Korean Research Councils Discuss Bridging Science, Humanities, and National Policy")
[Science Bulletin] Korean Research Councils Discuss Bridging Science, Humanities, and National Policy
동아사이언스 · 2026.04.12
MAHA and science-based public health: Can’t we all just get along? | Science-Based Medicine
Science-Based Medicine · 2026.04.15
Health Sciences Weekend Unites Alumni Around Shared Health Science Center Story - UTHSC News
UTHSC News · 2026.04.14
Molecular Sciences Building - Premier Construction
ROMA Publications Ltd · 2026.04.14
Monmouth College Enhances Computer Science Program
River Cities’ Reader · 2026.04.14
Designing Winners - College of Arts and Sciences
College of Arts and Sciences · 2026.04.14