Since 2000 a number of controversial research projects have raised concerns that states or terrorist groups might use published research results to replicate the work for malevolent purposes. The issue came into greater focus in December 2011, after it was revealed that the U.S.-based National Science Advisory Board for Biosecurity had requested one U.S. and another Dutch lab to halt publication of work describing what was done to enhance the transmissibility of the H5N1 flu virus to mammals. In spite of a decade of debate and policy development to prevent bio-attacks and the misuse of dual-use research, little effort has been made to determine whether scientific publications can indeed help replication by untrained individuals. This paper taps into the extensive science and technology literature and recent studies in the security field, to show that replication of scientific work is inherently difficult and probably inaccessible to untrained individuals. The paper also offers new policy options designed to more effectively prevent the misuse of dual-use research.