Aug 20, 2026 Chemical Safety: From Toxic Ignorance to Intelligent Testing
The Dilemma
In July 1997, the Environmental Defense Fund (EDF) published a study, Toxic Ignorance, noting that 71% of high-production-volume (HPV) chemicals produced by industry in quantities of 1 million pounds or more per year lacked adequate safety data. The subsequent negotiations with chemical companies led to an agreement to speed up testing to fill those data gaps.
However, animal advocates were concerned that expanded testing would dramatically increase the use of laboratory animals. For nearly three decades, environmental and animal protection organizations have wrestled with how to improve chemical safety information while minimizing animal testing.
Today, the testing landscape is changing. The information available on chemicals in commerce has improved significantly, and costly animal tests are increasingly being supplemented or replaced by faster, more efficient approaches that provide relevant human safety data.
The Scale of the Challenge: How Many Chemicals are There?
This simple question turns out to be much more complex than the reader might assume. Estimates of the total number of chemicals in the world range from 1018 to 10200. These are huge numbers.
However, a recent study reports that the number of chemicals and chemical mixtures registered for commercial production and use is “only” 350,000, and includes 157,000 identified chemicals, 75,000 chemical mixtures and polymers, and 120,000 substances that could not be uniquely identified.
The estimate of 350,000 registered chemicals is much larger than an earlier 2019 estimate by Gregory Bond and Veronique Garny of 40-60,000 industrial chemicals in global commerce. However, Bond and Garny estimated that only around 6,000 of those chemicals were thought to account for more than 99% of the total volume of chemicals in commerce globally.
REACH and Chemical Reporting
REACH, the European Union’s 2007 chemical safety legislation, requires different levels of safety reporting depending on the quantity of a chemical produced or imported. Chemicals manufactured in very large volumes (more than 1,000 tons per annum) require extensive toxicity data and risk assessments, while lower-volume chemicals face less demanding reporting requirements.
Progress on the EDF Agreement with Chemical Manufacturers
Also in 2007, the EDF revisited its agreement with chemical manufacturers to produce basic toxicity data on high-production-volume chemicals. According to the EDF, manufacturers had provided data on a little more than half of the 2,800 HPV chemicals. Reportedly, the U.S. EPA had exempted or removed about 15% of the original list of HPV chemicals, and industry sponsors could not be identified for 10% of the chemicals, while industry representatives argued that companies had invested significant resources in a good-faith effort to generate the requested safety information.
Traditional Chemical Tests Fall Short
Historically, the safety testing of industrial chemicals has relied heavily on studies involving laboratory animals, typically rats, rabbits, dogs or primates. Such tests are time-consuming and expensive, and do not always (or usually?) generate safety information relevant to human exposure.
Safety testing capacity is also limited. One expert estimated that the world could only conduct a full range of safety tests on around 50 chemicals a year. Given the enormous number of chemicals in commerce and the existing data gaps, conventional animal-based testing cannot realistically be expected to reduce them.
New Approaches
Fortunately, new technologies are being developed that can assess chemical safety by combining existing data with artificial intelligence systems to identify problematic chemicals.
Drs. Nicole Kleinstreuer of the U.S. National Institutes of Health and Thomas Hartung of the Johns Hopkins Center for Alternatives to Animal Testing have published a series of papers (e.g., see Altex for a recent example) on the potential of modern Artificial Intelligence tools to produce accurate predictions of the toxicity of modern chemicals very quickly and cheaply.
When additional validation is needed, those predictions can be supplemented with New Approach Methods (NAMs), including automated systems based on human cell cultures. Together, these innovations provide a faster, more effective, more realistic, and more humane path to chemical safety assessment.
Conclusion
The debate that began with Toxic Ignorance nearly thirty years ago remains relevant. Society needs reliable information about the safety of hundreds of thousands of chemicals in commercial use. Traditional animal testing is not the solution. New approach methods (NAMs) and Artificial Intelligence, developed by an increasing number of academic, corporate, and government research teams, are building the foundation for much quicker and more efficient chemical safety and risk assessment protocols.
Video Credit: Fitztudio, iStock