[en] After having highlighted the main difficulties met when trying to take decisions on emission reductions (i.e. disequilibrium between short term action and long term dynamics, series of uncertainties, lack of a clear information on threats), this research aims at studying the extent and the agenda of CO₂ emission reduction efforts at a global scale. First, the author explores under various perspectives the relationship between the extent of climate change and the extent of its impacts and modes of analysis of climate policies. Then, he examines how different attitudes towards the climate risk are translated into recommendations for action which are obtained by always more complex models. In order to compare results, simulations are performed, based on optimal control models belonging to the RESPONSE family. Three different options or attitudes are then studied and assessed with RESPONSE: firstly, the uncertainty on impact and damage assessment is still too important to have confidence in present assessments; secondly, while being sceptical on damage assessment, more confidence is given to climate model results, and thirdly, an explicit representation of climate damages is used to introduce rationality issues in the debate on risks