In real life, the perception of a risk depends not only on its reality and severity, but also on subjective external factors. The fear of potential health risks of ionizing radiation is very present in collective unconscious of our society, most probably because of their military use, which preceded their civil and medical use. This, as well as the principle of the risk/benefit balance optimization inherent to any good medical practice, makes this topic of major importance for the diffusion of nuclear medicine. The therapeutic field uses high doses were the risk is always present, but with a large expected benefit. In contrast, imaging field is in the range of very low doses where the risk is still unknown. But for both applications, major improvements have been and still can be done, and this supposes for each nuclear physician or physicist to be aware and well trained in the field. Potential health risks of ionizing radiation were identified soon after the discovery of x-rays in 1895. In the field of Nuclear Medicine, functional and molecular imaging as well as therapy demonstrated their benefits in terms of medical usefulness. Medical diagnostic exposure is nowadays the first cause of radiation exposure, raising some concerns regarding potential side effects, which are to be weighed against the specific risk of the disease(s) that are taken in charge, mainly in elderly patients. However, optimization continuously aims to decrease these side effects for the same or extended benefit [1]. Injuries to highly exposed tissues, such as encountered in therapeutic applications of Nuclear Medicine, are classified as deterministic effects because they will always occur once a particular threshold dose has been exceeded. Tissue reactions are caused by cell death and increase in severity as radiation dose increases. Deterministic effects can impair the tissue integrity, and compromise the function of non-target organs. A threshold dose is needed for damage to become clinically observable, and the extent of damage depends upon the absorbed dose, dose rate, and radiation quality. Thus, the severity of the effect increases with increasing absorbed dose. Apart from the case of accidental irradiation with early clinical reaction, these deterministic effects are usually delayed, occurring years or sometimes decades later and include tissue reactions such as cataracts, cardiovascular disorders, and necrosis. The risk of stochastic effects increases with dose, in theory with no threshold. In addition to hereditary effects, which originate in germline mutations, cancer is classified as a stochastic effect since it originates in somatic mutations. Even if the reality of this risk is still uncertain in the domain of very low doses such as used in the field of imaging, this has led to the adoption of international rules in the field. Cancer risk increases after exposure to moderate and high doses of radiation (more than 0.1–0.2 Gy); however, whether cancer risk is increased below this level is unclear and is the object of extensive research. Currently, one in three women and one in two men will develop cancer in his or her lifetime. Since this is so much higher than any estimated effect of medical radiation, and since medical exposure is in the same range than natural exposure, the potential increase in cancer due to radiation is extremely difficult to detect. In addition, no prospective Dominique Le Guludec and Jocelyne Aigueperse contributed equally to the literature search and manuscript writing.