The emergence of web 2.0 social media enables the gradual emergence of a second generation of e-participation characterized by more citizens’ control, in which government agencies post content (e.g. short or longer text, images, video) to various social media and then analyze citizens’ interactions with it (e.g. views, likes/dislikes, comments, etc.). In this paper we propose an even more citizens controlled third generation of e-participation exploiting web 2.0 social media as well, but in a different manner. It is based on the search by government agencies for content on a public policy under formulation, which has been created in a large set of web 2.0 sources (e.g. blogs and microblogs, news sharing sites, online forums) by citizens freely, without any initiation, stimulation or moderation through government postings. This content undergoes advanced processing in order to extract from it arguments, opinions, issues and proposals on the particular policy, identify their sentiments (positive or negative), and finally summarize and visualize them. This approach allows the exploitation of the vast amount of user-generated content created in numerous web 2.0 social media for supporting governments to understand better the needs, wishes and beliefs of citizens, and create better and more socially rooted policies.