Risky situations are situations mostly characterized by uncertainty. One essential aim of preventive actions is actually to minimize this uncertainty and to make these situations more or less foreseable, more or less controllable. Risk appraisal and accident causal explanations are basic prequisites for risk prevention. Because of the high level of uncertainty in risky situations, risk assessment and accident causal explanation are inevitably dominated by the representations and beliefs that hold the different actors facing them. In fact, risky situations are generally complex?; not only this complexity contributes to increase the level of uncertainty, but it also generates beliefs likely to lead to biased or erroneous causal inferences and conclusions. Such biases are likely to generate false certainties (or beliefs). The present article starts with a description of the link between the notions of risk and uncertainty by stressing the fact that uncertainty is an important component of risk appraisal. We then show that what is called risk is in fact a cognitive and social construct. Each Society and each individual in a Society according to the epochs could have their own definition of the risk they consider to be tolerable or not, controllable or not, beneficial or harmful. Next, we show that risk appraisal and accident explanation are opportunities particularly favourable to the arousal of beliefs and causal inferences. The role of such beliefs and inferences is to fulfill the gap in scientific knowledge regarding risky situations. In face of this gap, people tend to use some heuristics (see for example Kahneman, Slovic, &?Tversky, 1982) in order to reduce the uncertainty and to facilitate risk assessment. But many factors (cognitive, motivational, affective) could lead to distortions in risk perception and accident explanation. Some of the factors influencing risk appraisal or risk perception and accident explanation are presented. They are either linked to the characteristics of the risk itself or to the characteristics proper to the perceiver or the explainer. These factors lead to some illusions and false certainties in risk perception and accident explanation. Such false beliefs or illusions could lead to unsafe behaviours. Finally, we show that investigating these illusory beliefs could be useful to design relevant preventive messages and to influence more efficiently individual behaviours in favour of safety.