Abstract: Risk assessments have several sources of uncertainty, including external and internal exposure, potential metabolites, the nature of the dose-response curve, mode-of-action, population variability and when extrapolating from animals to humans. In the past, risk assessors have used uncertainty factors to calculate risk in the face of these data, but a realistic and accurate assessment of uncertainty can be avoided depending on such worst-case scenarios. This may allow less conservative safety margins or alternatively a better understanding of the most precise way to ensure a chemical is used safely. Chemoinformatics, predictive modelling and big data approaches can be used to significantly reduce uncertainty about many key aspects of hazard identification and exposure, and by focusing on human models avoid uncertainty introduced by depending on animals. Moreover, employing these tools can aid in pinpointing the optimal testing strategy that minimizes uncertainty to the greatest extent and as efficiently as possible.