Are the chemicals around you safe? Researchers are using AI to find out – Texas A&M Stories

Researchers at Texas A&M University are utilizing artificial intelligence (AI) to address the gap in chemical toxicity testing due to the vast number of unstudied chemicals.
Dr. Weihsueh Chiu, from Texas A&M’s College of Veterinary Medicine and Biomedical Sciences, leads the development of AI tools that predict chemical toxicity and assess the reliability of these predictions.
Traditional toxicity tests, which rely on animal studies or human epidemiological research, are time-consuming, costly, and limited in scope, leading to a significant gap in chemical safety data.
The AI tools developed by Chiu’s team employ a two-stage machine learning framework that uses real-world chemical properties instead of abstract descriptors, making predictions more interpretable and trustworthy.
The introduction of “uncertainty-aware” machine learning allows the models to estimate the reliability of each prediction, guiding where more research is needed based on varying levels of uncertainty.
These AI tools generate ranges of possible outcomes for each prediction, helping to identify potential risks more effectively and shifting focus from a reactive to a proactive approach in chemical safety assessments.
The ultimate goal is to allow regulators to determine which substances require stricter regulation, more testing, or removal from the market based on these AI insights.

You may have missed