Vignollet, L.Plu, M.Marty, J. C.Agosto, L.2017-04-152017-04-152005978-1-4020-3591-3This paper presents how an information exchange network can be improved by users' collaboration. This social media is based on content recommendation. Instead of using an automated content recommender system, we suggest an alternative approach where the information comes from trusted users. In order to overcome traditional problems of an open social media, we propose some regulation mechanisms. First each user manually controls her/his contacts network. Second we have introduced a contact recommender system to help users to carefully open their closed relationship network. This recommender system selects the recommended relationships in such a way it should optimize some global qualities of the social media. This paper details the algorithms of this recommendation process.enText10.1007/1-4020-3591-8_22