08-10-2013
Professor Luís Moniz Pereira from FCT-UNL's Computer Science Department published a work under the title"Good Agreements make good friends" in the prestigious NATURE's open access journal Scientific Reports that might have direct incidence for robot morality. The premise of this work lies under the assumption according to which conventional wisdom suggests that clear agreements need to be made prior to any collaborative effort in order to avoid potential frustrations for the participants during or at the end of a joint project. In a mathematical and numerical investigation The Anh Han and Tom Lenaerts (both from VUB and ULB) show, in collaboration with Luís Moniz Pereira and Francisco C. Santos from Portugal, that this behavior may actually have been shaped by natural selection. Moreover, their analysis shows that making prior agreements on the consequences of not honoring a deal provide a more effective road to cooperation than simply punishing bad behavior after the facts, even when there is a cost associated to setting up the agreements. This work can be read in http://www.nature.com/srep/2013/130918/srep02695/full/srep02695.html .
