It is amazing to see how much the approach of the food risk analysis evolved in the recent years. For half a century and the birth of the risk assessment methodology in the food domain, only no appreciable health risk was considered acceptable by the manager. This is the vocabulary used in the case of a voluntary, deliberated human action, as the use of food additives (definition of ADI). In the case of risks not resulting from such an action, as that of the presence of contaminants, the risk assessor allocates provisional tolerable daily, weekly or monthly intake that are the basis for regulation. This vocabulary is in agreement with the objective which consists in approaching closer possible of the zero risk which is the wish of a majority of the consumers. Some years ago, the risk managers insisted to obtain from the assessors as often as possible a quantitative risk evaluation. More recently even, the managers would like to decide on the basis of a balance of risk and benefit acceptable for management purposes. Finally, they hope that general principles and tools will be available for conducting a quantitative risk-benefit analysis for foods and food ingredients. What is possible in the case of functional foods (FF)? Based on the definition of FF proposed in the programme FUFOSE, one has to distinguish between different situations in order to assess the risk: that of a micro-, that of a macro-component or that of a whole food. These situations have been clearly described in the document resulting from FOSIE. The standardized methodology relevant to assess micro-components is not well adapted to the assessment of whole food. Concepts of substantial equivalence and of history of safe use could be useful tools in this case. However, quantitative risk assessment remains a very difficult exercise. If a process for the assessment of health benefit of FF has been proposed as an outcome of the PASSCLAIM action, the quantification of this benefit needs adequate tools. An EFSA scientific colloquium on “Risk-Benefit Analysis of Foods” organized in July 2006 concluded that the risk-benefit analysis should mirror the current risk analysis paradigm and that its assessment should be performed with common scales. Disability adjusted life years (DALYs) or quality adjusted life years (QUALYs) have been proposed as some of these common scales. However, the meeting “concluded that the data available to undertake a quantitative risk-benefit assessment may be too scarce”. Because it was considered that it was premature to formulate guidelines on good risk-benefit analysis practice and it is now time to “learning by doing”, a reference to the upcoming ILSI Europe project BRAFO was done. All these aspects are discussed, in particular in relation to the specific case of FF.