The document discusses the increasing use of computational models for predicting human toxicities in pharmaceutical development due to new regulations requiring toxicity data. It highlights the feasibility of predicting toxicity from molecular structures and mentions collaborative efforts and technological advancements that facilitate the development of predictive models and mobile applications. Examples of computational models for specific toxicities, such as drug-induced liver injury, are provided, emphasizing their potential to improve drug safety and reduce reliance on animal testing.