My understanding of liposomes as vectors is that they are less efficient than viruses and their robustness is lower than viral vectors. One needs to make sure that the plasmid (or plasmids) gets out of the liposome and into the target cell without getting digested along the way. Viruses, as we know, are quite good at getting DNA (or RNA) into living cells. I'm sure people are working on ways to extend their lifetime in the body, but I recall hearing that liposomes get scavenged by the body pretty quickly. Liposomal methods mostly been used for direct injection into a target area, as a result, which thus far limits their applicability. They may not be anywhere near as much of a direct physical health risk as viruses, but trying to get them to work as well as viruses in certain regards may take a toll on a researcher's mental health. Heh. Naked DNA shouldn't be much of a problem (we have nucleases on our skin), and the others would probably vary depending on the details.
One still, of course, has the standard issues of concerns with gene therapy regardless of delivery method (making sure that it incorporates into the genome in a beneficial manner, that it doesn't spark an immunological response against the gene therapy-treated cell, and so on).