A threshold represents the theoretically defined dose, below which no abnormal increase in response is observed. Genotoxic carcinogens are known to have an irreversible effect on the genetic cellular structure. Even in small amounts, genotoxic carcinogens are assumed to have additive effects and therefore to subject individuals exposed to them to an incremental risk of developing cancer. Based on this assumption, the no-threshold concept was introduced exclusively for genotoxic carcinogens and has been adopted in Japan as a basis for the regulatory risk assessment in case of such chemicals. The current regulatory policy adheres to the fundamental principle of food safety i.e. to the precedence of protecting people's health. Dose-response studies recently conducted on various genotoxic agents suggest the existence of a threshold. If confirmed, such findings may provide sufficient scientific evidence to substantiate the adoption of a threshold concept as the principle of the regulatory assessment of the risks of genotoxic carcinogens and their impact on health. It should however be emphasized that the limitations of a threshold approach must be clearly understood and presented to lend credence to the proposition of a paradigm shift from the current regulatory policy: A threshold is not a value that can be determined directly and precisely from dose-response data, but one that can be estimated from analytical data by means of a logically-elaborated mathematical model calculation. Scientific efforts in support of the adoption of a threshold in this context should therefore be focused on the development of appropriate mathematical models, and on the establishment of toxicological concepts that substantiate their application. A realistic first step towards a paradigm shift from the no-threshold concept is to seek general consensus on the introduction of an appropriately estimated“virtually safe dose”, instead of a threshold.